- title: 'Conference on Learning Theory 2020: Preface'
abstract: 'Preface to the proceedings of the 32nd Conference on Learning Theory'
volume: 125
URL: http://proceedings.mlr.press/v125/abernethy20a.html
PDF: http://proceedings.mlr.press/v125/abernethy20a/abernethy20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-abernethy20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 1-2
id: abernethy20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 1
lastpage: 2
published: 2020-07-15 00:00:00 +0000
- title: 'Domain Compression and its Application to Randomness-Optimal Distributed Goodness-of-Fit'
abstract: ' We study goodness-of-fit of discrete distributions in the distributed setting, where samples are divided between multiple users who can only release a limited amount of information about their samples due to various information constraints. Recently, a subset of the authors showed that having access to a common random seed (i.e., shared randomness) leads to a significant reduction in the sample complexity of this problem. In this work, we provide a complete understanding of the interplay between the amount of shared randomness available, the stringency of information constraints, and the sample complexity of the testing problem by characterizing a tight trade-off between these three parameters. We provide a general distributed goodness-of-fit protocol that as a function of the amount of shared randomness interpolates smoothly between the private- and public-coin sample complexities. We complement our upper bound with a general framework to prove lower bounds on the sample complexity of this testing problems under limited shared randomness. Finally, we instantiate our bounds for the two archetypal information constraints of communication and local privacy, and show that our sample complexity bounds are optimal as a function of all the parameters of the problem, including the amount of shared randomness. A key component of our upper bounds is a new primitive of \textit{domain compression}, a tool that allows us to map distributions to a much smaller domain size while preserving their pairwise distances, using a limited amount of randomness. '
volume: 125
URL: http://proceedings.mlr.press/v125/acharya20a.html
PDF: http://proceedings.mlr.press/v125/acharya20a/acharya20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-acharya20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Acharya
given: Jayadev
- family: Canonne
given: Clément L
- family: Han
given: Yanjun
- family: Sun
given: Ziteng
- family: Tyagi
given: Himanshu
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 3-40
id: acharya20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 3
lastpage: 40
published: 2020-07-15 00:00:00 +0000
- title: 'Distributed Signal Detection under Communication Constraints'
abstract: ' Independent draws from a $d$-dimensional spherical Gaussian distribution are distributed across users, each holding one sample. A central server seeks to distinguish between the two hypotheses: the distribution has zero mean, or the mean has $\ell_2$-norm at least $\varepsilon$, a pre-specified threshold. However, the users can each transmit at most $\ell$ bits to the server. This is the problem of detecting whether an observed signal is simply white noise in a distributed setting. We study this distributed testing problem with and without the availability of a common randomness shared by the users. We design schemes with and without such shared randomness which achieve sample complexities. We then obtain lower bounds for protocols with public randomness, tight when $\ell=O(1)$. We finally conclude with several conjectures and open problems.'
volume: 125
URL: http://proceedings.mlr.press/v125/acharya20b.html
PDF: http://proceedings.mlr.press/v125/acharya20b/acharya20b.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-acharya20b.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Acharya
given: Jayadev
- family: Canonne
given: Clément L
- family: Tyagi
given: Himanshu
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 41-63
id: acharya20b
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 41
lastpage: 63
published: 2020-07-15 00:00:00 +0000
- title: 'Optimality and Approximation with Policy Gradient Methods in Markov Decision Processes'
abstract: ' Policy gradient (PG) methods are among the most effective methods in challenging reinforcement learning problems with large state and/or action spaces. However, little is known about even their most basic theoretical convergence properties, including: if and how fast they converge to a globally optimal solution (say with a sufficiently rich policy class); how they cope with approximation error due to using a restricted class of parametric policies; or their finite sample behavior. Such characterizations are important not only to compare these methods to their approximate value function counterparts (where such issues are relatively well understood, at least in the worst case), but also to help with more principled approaches to algorithm design. This work provides provable characterizations of computational, approximation, and sample size issues with regards to policy gradient methods in the context of discounted Markov Decision Processes (MDPs). We focus on both: 1) “tabular” policy parameterizations, where the optimal policy is contained in the class and where we show global convergence to the optimal policy, and 2) restricted policy classes, which may not contain the optimal policy and where we provide agnostic learning results. In the \emph{tabular setting}, our main results are: 1) convergence rate to global optimum for direct parameterization and projected gradient ascent 2) an asymptotic convergence to global optimum for softmax policy parameterization and PG; and a convergence rate with additional entropy regularization, and 3) dimension-free convergence to global optimum for softmax policy parameterization and Natural Policy Gradient (NPG) method with exact gradients. In \emph{function approximation}, we further analyze NPG with exact as well as inexact gradients under certain smoothness assumptions on the policy parameterization and establish rates of convergence in terms of the quality of the initial state distribution. One insight of this work is in formalizing how a favorable initial state distribution provides a means to circumvent worst-case exploration issues. Overall, these results place PG methods under a solid theoretical footing, analogous to the global convergence guarantees of iterative value function based algorithms.'
volume: 125
URL: http://proceedings.mlr.press/v125/agarwal20a.html
PDF: http://proceedings.mlr.press/v125/agarwal20a/agarwal20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-agarwal20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Agarwal
given: Alekh
- family: Kakade
given: Sham M
- family: Lee
given: Jason D
- family: Mahajan
given: Gaurav
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 64-66
id: agarwal20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 64
lastpage: 66
published: 2020-07-15 00:00:00 +0000
- title: 'Model-Based Reinforcement Learning with a Generative Model is Minimax Optimal'
abstract: ' This work considers the sample and computational complexity of obtaining an $\epsilon$-optimal policy in a discounted Markov Decision Process (MDP), given only access to a generative model. In this model, the learner accesses the underlying transition model via a sampling oracle that provides a sample of the next state, when given any state-action pair as input. We are interested in a basic and unresolved question in model based planning: is this naïve “plug-in” approach — where we build the maximum likelihood estimate of the transition model in the MDP from observations and then find an optimal policy in this empirical MDP — non-asymptotically, minimax optimal? Our main result answers this question positively. With regards to computation, our result provides a simpler approach towards minimax optimal planning: in comparison to prior model-free results, we show that using \emph{any} high accuracy, black-box planning oracle in the empirical model suffices to obtain the minimax error rate. The key proof technique uses a leave-one-out analysis, in a novel “absorbing MDP” construction, to decouple the statistical dependency issues that arise in the analysis of model-based planning; this construction may be helpful more generally.'
volume: 125
URL: http://proceedings.mlr.press/v125/agarwal20b.html
PDF: http://proceedings.mlr.press/v125/agarwal20b/agarwal20b.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-agarwal20b.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Agarwal
given: Alekh
- family: Kakade
given: Sham
- family: Yang
given: Lin F.
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 67-83
id: agarwal20b
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 67
lastpage: 83
published: 2020-07-15 00:00:00 +0000
- title: 'From Nesterov’s Estimate Sequence to Riemannian Acceleration'
abstract: ' We propose the first global accelerated gradient method for Riemannian manifolds. Toward establishing our results, we revisit Nesterov’s estimate sequence technique and develop a conceptually simple alternative from first principles. We then extend our analysis to Riemannian acceleration, localizing the key difficulty into “metric distortion.” We control this distortion via a novel geometric inequality, which enables us to formulate and analyze global Riemannian acceleration.'
volume: 125
URL: http://proceedings.mlr.press/v125/ahn20a.html
PDF: http://proceedings.mlr.press/v125/ahn20a/ahn20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-ahn20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Ahn
given: Kwangjun
- family: Sra
given: Suvrit
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 84-118
id: ahn20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 84
lastpage: 118
published: 2020-07-15 00:00:00 +0000
- title: 'Closure Properties for Private Classification and Online Prediction'
abstract: ' Let H be a class of boolean functions and consider a composed class H’ that is derived from H using some arbitrary aggregation rule (for example, H’ may be the class of all 3-wise majority-votes of functions in H). We upper bound the Littlestone dimension of H’ in terms of that of H. As a corollary, we derive closure properties for online learning and private PAC learning. The derived bounds on the Littlestone dimension exhibit an undesirable exponential dependence. For private learning, we prove close to optimal bounds that circumvents this suboptimal dependency. The improved bounds on the sample complexity of private learning are derived algorithmically via transforming a private learner for the original class H to a private learner for the composed class H’. Using the same ideas we show that any (proper or improper) private algorithm that learns a class of functions H in the realizable case (i.e., when the examples are labeled by some function in the class) can be transformed to a private algorithm that learns the class H in the agnostic case.'
volume: 125
URL: http://proceedings.mlr.press/v125/alon20a.html
PDF: http://proceedings.mlr.press/v125/alon20a/alon20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-alon20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Alon
given: Noga
- family: Beimel
given: Amos
- family: Moran
given: Shay
- family: Stemmer
given: Uri
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 119-152
id: alon20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 119
lastpage: 152
published: 2020-07-15 00:00:00 +0000
- title: 'Hierarchical Clustering: A 0.585 Revenue Approximation'
abstract: ' Hierarchical Clustering trees have been widely accepted as a useful form of clustering data, resulting in a prevalence of adopting fields including phylogenetics, image analysis, bioinformatics and more. Recently, Dasgupta (STOC 16’) initiated the analysis of these types of algorithms through the lenses of approximation. Later, the dual problem was considered by Moseley and Wang (NIPS 17’) dubbing it the Revenue goal function. In this problem, given a nonnegative weight $w_{ij}$ for each pair $i,j \in [n]=\{1,2, \ldots ,n\}$, the objective is to find a tree $T$ whose set of leaves is $[n]$ that maximizes the function $\sum_{i1$; (2) almost exact recovery for $ 1 \le k \le o\left( \frac{\log n}{\log \log n} \right)$ if $ \liminf \frac{kD(P_n||Q_n)}{\log n}>1, $ where $\alpha_n \triangleq -2 \log \int \sqrt{d P_n d Q_n}$ is the Rényi divergence of order $\frac{1}{2}$ and $D(P_n||Q_n)$ is the Kullback-Leibler divergence. Under mild distributional assumptions, these conditions are shown to be information-theoretically necessary for any algorithm to succeed. A key challenge in the analysis is the enumeration of $2k$-NN graphs that differ from the hidden one by a given number of edges. We also analyze several computationally efficient algorithms and provide sufficient conditions under which they achieve exact/almost exact recovery. In particular, we develop a polynomial-time algorithm that attains the threshold for exact recovery under the small-world model.'
volume: 125
URL: http://proceedings.mlr.press/v125/ding20a.html
PDF: http://proceedings.mlr.press/v125/ding20a/ding20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-ding20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Ding
given: Jian
- family: Wu
given: Yihong
- family: Xu
given: Jiaming
- family: Yang
given: Dana
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 1540-1553
id: ding20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 1540
lastpage: 1553
published: 2020-07-15 00:00:00 +0000
- title: 'Root-n-Regret for Learning in Markov Decision Processes with Function Approximation and Low Bellman Rank'
abstract: ' In this paper, we consider the problem of online learning of Markov decision processes (MDPs) with very large state spaces. Under the assumptions of realizable function approximation and low Bellman ranks, we develop an online learning algorithm that learns the optimal value function while at the same time achieving very low cumulative regret during the learning process. Our learning algorithm, Adaptive Value-function Elimination (AVE), is inspired by the policy elimination algorithm proposed in (Jiang et al., 2017), known as OLIVE. One of our key technical contributions in AVE is to formulate the elimination steps in OLIVE as contextual bandit problems. This technique enables us to apply the active elimination and expert weighting methods from (Dudik et al., 2011), instead of the random action exploration scheme used in the original OLIVE algorithm, for more efficient exploration and better control of the regret incurred in each policy elimination step. To the best of our knowledge, this is the first root-n-regret result for reinforcement learning in stochastic MDPs with general value function approximation.'
volume: 125
URL: http://proceedings.mlr.press/v125/dong20a.html
PDF: http://proceedings.mlr.press/v125/dong20a/dong20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-dong20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Dong
given: Kefan
- family: Peng
given: Jian
- family: Wang
given: Yining
- family: Zhou
given: Yuan
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 1554-1557
id: dong20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 1554
lastpage: 1557
published: 2020-07-15 00:00:00 +0000
- title: 'Embedding Dimension of Polyhedral Losses'
abstract: ' A common technique in supervised learning with discrete losses, such as 0-1 loss, is to optimize a convex surrogate loss over Rd, calibrated with respect to the original loss. In particular, recent work has investigated embedding the original predictions (e.g. labels) as points in Rd, showing an equivalence to using polyhedral surrogates. In this work, we study the notion of the embedding dimension of a given discrete loss: the minimum dimension d such that an embedding exists. We characterize d-embeddability for all d, with a particularly tight characterization for d=1 (embedding into the real line), and useful necessary conditions for d>1 in the form of a quadratic feasibility program. We illustrate our results with novel lower bounds for abstain loss.'
volume: 125
URL: http://proceedings.mlr.press/v125/finocchiaro20a.html
PDF: http://proceedings.mlr.press/v125/finocchiaro20a/finocchiaro20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-finocchiaro20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Finocchiaro
given: Jessie
- family: Frongillo
given: Rafael
- family: Waggoner
given: Bo
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 1558-1585
id: finocchiaro20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 1558
lastpage: 1585
published: 2020-07-15 00:00:00 +0000
- title: 'Efficient Parameter Estimation of Truncated Boolean Product Distributions'
abstract: ' We study the problem of estimating the parameters of a Boolean product distribution in $d$ dimensions, when the samples are truncated by a set $S \subset \{0, 1\}^d$ accessible through a membership oracle. This is the first time that the computational and statistical complexity of learning from truncated samples is considered in a discrete setting. We introduce a natural notion of \emph{fatness} of the truncation set $S$, under which truncated samples reveal enough information about the true distribution. We show that if the truncation set is sufficiently fat, samples from the true distribution can be generated from truncated samples. A stunning consequence is that virtually any statistical task (e.g., learning in total variation distance, parameter estimation, uniformity or identity testing) that can be performed efficiently for Boolean product distributions, can also be performed from truncated samples, with a small increase in sample complexity. We generalize our approach to ranking distributions over $d$ alternatives, where we show how fatness implies efficient parameter estimation of Mallows models from truncated samples. Exploring the limits of learning discrete models from truncated samples, we identify three natural conditions that are necessary for efficient identifiability: (i) the truncation set $S$ should be rich enough; (ii) $S$ should be accessible through membership queries; and (iii) the truncation by $S$ should leave enough randomness in all directions. By carefully adapting the Stochastic Gradient Descent approach of (Daskalakis et al., FOCS 2018), we show that these conditions are also sufficient for efficient learning of truncated Boolean product distributions.'
volume: 125
URL: http://proceedings.mlr.press/v125/fotakis20a.html
PDF: http://proceedings.mlr.press/v125/fotakis20a/fotakis20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-fotakis20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Fotakis
given: Dimitris
- family: Kalavasis
given: Alkis
- family: Tzamos
given: Christos
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 1586-1600
id: fotakis20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 1586
lastpage: 1600
published: 2020-07-15 00:00:00 +0000
- title: 'Rigorous Guarantees for Tyler’s M-Estimator via Quantum Expansion'
abstract: ' Estimating the shape of an elliptical distribution is a fundamental problem in statistics. One estimator for the shape matrix, Tyler’s M-estimator, has been shown to have many appealing asymptotic properties. It performs well in numerical experiments and can be quickly computed in practice by a simple iterative procedure. Despite the many years the estimator has been studied in the statistics community, there was neither a non-asymptotic bound on the rate of the estimator nor a proof that the iterative procedure converges in polynomially many steps. Here we observe a surprising connection between Tyler’s M-estimator and operator scaling, which has been intensively studied in recent years in part because of its connections to the Brascamp-Lieb inequality in analysis. We use this connection, together with novel results on quantum expanders, to show that Tyler’s M-estimator has the optimal rate up to factors logarithmic in the dimension, and that in the generative model the iterative procedure has a linear convergence rate even without regularization.'
volume: 125
URL: http://proceedings.mlr.press/v125/franks20a.html
PDF: http://proceedings.mlr.press/v125/franks20a/franks20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-franks20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Franks
given: William Cole
- family: Moitra
given: Ankur
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 1601-1632
id: franks20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 1601
lastpage: 1632
published: 2020-07-15 00:00:00 +0000
- title: 'From tree matching to sparse graph alignment'
abstract: ' In this paper we consider alignment of sparse graphs, for which we introduce the Neighborhood Tree Matching Algorithm (NTMA). For correlated Erdős-R{é}nyi random graphs, we prove that the algorithm returns – in polynomial time – a positive fraction of correctly matched vertices, and a vanishing fraction of mismatches. This result holds with average degree of the graphs in $O(1)$ and correlation parameter $s$ that can be bounded away from $1$, conditions under which random graph alignment is particularly challenging. As a byproduct of the analysis we introduce a matching metric between trees and characterize it for several models of correlated random trees. These results may be of independent interest, yielding for instance efficient tests for determining whether two random trees are correlated or independent.'
volume: 125
URL: http://proceedings.mlr.press/v125/ganassali20a.html
PDF: http://proceedings.mlr.press/v125/ganassali20a/ganassali20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-ganassali20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Ganassali
given: Luca
- family: Massoulié
given: Laurent
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 1633-1665
id: ganassali20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 1633
lastpage: 1665
published: 2020-07-15 00:00:00 +0000
- title: 'On the Convergence of Stochastic Gradient Descent with Low-Rank Projections for Convex Low-Rank Matrix Problems'
abstract: ' We revisit the use of Stochastic Gradient Descent (SGD) for solving convex optimization problems that serve as highly popular convex relaxations for many important low-rank matrix recovery problems such as matrix completion, phase retrieval, and more. The computational limitation of applying SGD to solving these relaxations in large-scale is the need to compute a potentially high-rank singular value decomposition (SVD) on each iteration in order to enforce the low-rank-promoting constraint. We begin by considering a simple and natural sufficient condition so that these relaxations indeed admit low-rank solutions. This condition is also necessary for a certain notion of low-rank-robustness to hold. Our main result shows that under this condition which involves the eigenvalues of the gradient vector at optimal points, SGD with mini-batches, when initialized with a “warm-start" point, produces iterates that are low-rank with high probability, and hence only a low-rank SVD computation is required on each iteration. This suggests that SGD may indeed be practically applicable to solving large-scale convex relaxations of low-rank matrix recovery problems. Our theoretical results are accompanied with supporting preliminary empirical evidence. As a side benefit, our analysis is quite simple and short.'
volume: 125
URL: http://proceedings.mlr.press/v125/garber20a.html
PDF: http://proceedings.mlr.press/v125/garber20a/garber20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-garber20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Garber
given: Dan
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 1666-1681
id: garber20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 1666
lastpage: 1681
published: 2020-07-15 00:00:00 +0000
- title: 'Asymptotic Errors for High-Dimensional Convex Penalized Linear Regression beyond Gaussian Matrices'
abstract: ' We consider the problem of learning a coefficient vector $\bf x_0 \in \mathbb R^N$ from noisy linear observations $\mathbf{y} = \mathbf{F}{\mathbf{x}_{0}}+\mathbf{w} \in \mathbb R^M$ in high dimensional limit $M,N \to \infty$ with $\alpha \equiv M/N$ fixed. We provide a rigorous derivation of an explicit formula —first conjectured using heuristics method from statistical physics— for the asymptotic mean squared error obtained by penalized convex estimators such as the LASSO or the elastic net, for a sequence of very generic random matrix $\mathbf{F}$ corresponding to rotationally invariant data matrices of arbitrary spectrum. The proof is based on a convergence analysis of an oracle version of vector approximate message-passing (oracle-VAMP) and on the properties of its state evolution equations. Our method leverages on and highlights the link between vector approximate message-passing, Douglas-Rachford splitting and proximal descent algorithms, extending previous results obtained with i.i.d. matrices for a large class of problems. We illustrate our results on some concrete examples and show that even though they are asymptotic, our predictions agree remarkably well with numerics even for very moderate sizes.'
volume: 125
URL: http://proceedings.mlr.press/v125/gerbelot20a.html
PDF: http://proceedings.mlr.press/v125/gerbelot20a/gerbelot20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-gerbelot20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Gerbelot
given: Cédric
- family: Abbara
given: Alia
- family: Krzakala
given: Florent
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 1682-1713
id: gerbelot20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 1682
lastpage: 1713
published: 2020-07-15 00:00:00 +0000
- title: 'No-Regret Prediction in Marginally Stable Systems'
abstract: ' We consider the problem of online prediction in a marginally stable linear dynamical system subject to bounded adversarial or (non-isotropic) stochastic perturbations. This poses two challenges. Firstly, the system is in general unidentifiable, so recent and classical results on parameter recovery do not apply. Secondly, because we allow the system to be marginally stable, the state can grow polynomially with time; this causes standard regret bounds in online convex optimization to be vacuous. In spite of these challenges, we show that the online least-squares algorithm achieves sublinear regret (improvable to polylogarithmic in the stochastic setting), with polynomial dependence on the system’s parameters. This requires a refined regret analysis, including a structural lemma showing the current state of the system to be a small linear combination of past states, even if the state grows polynomially. By applying our techniques to learning an autoregressive filter, we also achieve logarithmic regret in the partially observed setting under Gaussian noise, with polynomial dependence on the memory of the associated Kalman filter.'
volume: 125
URL: http://proceedings.mlr.press/v125/ghai20a.html
PDF: http://proceedings.mlr.press/v125/ghai20a/ghai20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-ghai20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Ghai
given: Udaya
- family: Lee
given: Holden
- family: Singh
given: Karan
- family: Zhang
given: Cyril
- family: Zhang
given: Yi
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 1714-1757
id: ghai20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 1714
lastpage: 1757
published: 2020-07-15 00:00:00 +0000
- title: 'Last Iterate is Slower than Averaged Iterate in Smooth Convex-Concave Saddle Point Problems'
abstract: ' In this paper we study the smooth convex-concave saddle point problem. Specifically, we analyze the last iterate convergence properties of the Extragradient (EG) algorithm. It is well known that the ergodic (averaged) iterates of EG converge at a rate of $O(1/T)$ (Nemirovski, 2004). In this paper, we show that the last iterate of EG converges at a rate of $O(1/\sqrt{T})$. To the best of our knowledge, this is the first paper to provide a convergence rate guarantee for the last iterate of EG for the smooth convex-concave saddle point problem. Moreover, we show that this rate is tight by proving a lower bound of $\Omega(1/\sqrt{T})$ for the last iterate. This lower bound therefore shows a quadratic separation of the convergence rates of ergodic and last iterates in smooth convex-concave saddle point problems.'
volume: 125
URL: http://proceedings.mlr.press/v125/golowich20a.html
PDF: http://proceedings.mlr.press/v125/golowich20a/golowich20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-golowich20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Golowich
given: Noah
- family: Pattathil
given: Sarath
- family: Daskalakis
given: Constantinos
- family: Ozdaglar
given: Asuman
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 1758-1784
id: golowich20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 1758
lastpage: 1784
published: 2020-07-15 00:00:00 +0000
- title: 'Locally Private Hypothesis Selection'
abstract: ' We initiate the study of hypothesis selection under local differential privacy. Given samples from an unknown probability distribution $p$ and a set of $k$ probability distributions $\mathcal{Q}$, we aim to output, under the constraints of $\varepsilon$-differential privacy, a distribution from $\mathcal{Q}$ whose total variation distance to $p$ is comparable to the best such distribution. This is a generalization of the classic problem of $k$-wise simple hypothesis testing, which corresponds to when $p \in \mathcal{Q}$, and we wish to identify $p$. Absent privacy constraints, this problem requires $O(\log k)$ samples from $p$, and it was recently shown that the same complexity is achievable under (central) differential privacy. However, the naive approach to this problem under local differential privacy would require $\tilde O(k^2)$ samples. We first show that the constraint of local differential privacy incurs an exponential increase in cost: any algorithm for this problem requires at least $\Omega(k)$ samples. Second, for the special case of $k$-wise simple hypothesis testing, we provide a non-interactive algorithm which nearly matches this bound, requiring $\tilde O(k)$ samples. Finally, we provide sequentially interactive algorithms for the general case, requiring $\tilde O(k)$ samples and only $O(\log \log k)$ rounds of interactivity. Our algorithms are achieved through a reduction to maximum selection with adversarial comparators, a problem of independent interest for which we initiate study in the parallel setting. For this problem, we provide a family of algorithms for each number of allowed rounds of interaction $t$, as well as lower bounds showing that they are near-optimal for every $t$. Notably, our algorithms result in exponential improvements on the round complexity of previous methods.'
volume: 125
URL: http://proceedings.mlr.press/v125/gopi20a.html
PDF: http://proceedings.mlr.press/v125/gopi20a/gopi20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-gopi20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Gopi
given: Sivakanth
- family: Kamath
given: Gautam
- family: Kulkarni
given: Janardhan
- family: Nikolov
given: Aleksandar
- family: Wu
given: Zhiwei Steven
- family: Zhang
given: Huanyu
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 1785-1816
id: gopi20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 1785
lastpage: 1816
published: 2020-07-15 00:00:00 +0000
- title: 'Bessel Smoothing and Multi-Distribution Property Estimation'
abstract: ' We consider a basic problem in statistical learning: estimating properties of multiple discrete distributions. Denoting by $\Delta_k$ the standard simplex over $[k]:=\{0,1,\ldots, k\}$, a property of $d$ distributions is a mapping from $\Delta_k^d$ to $\mathbb R$. These properties include well-known distribution characteristics such as Shannon entropy and support size ($d=1$), and many important divergence measures between distributions ($d=2$). The primary problem being considered is to learn the property value of an $\emph{unknown}$ $d$-tuple of distributions from its sample. The study of such problems dates back to the works of Efron and Thisted (1976b); Thisted and Efron (1987); Good (1953b); Carlton (1969), and has been pushed forward steadily during the past decades. Surprisingly, before our work, the general landscape of this fundamental learning problem was insufficiently understood, and nearly all the existing results are for the special case $d\le 2$. Our first main result provides a near-linear-time computable algorithm that, given independent samples from any collection of distributions and for a broad class of multi-distribution properties, learns the property as well as the empirical plug-in estimator that uses samples with logarithmic-factor larger sizes. As a corollary of this, for any $\varepsilon>0$ and fixed $d\in \mathbb Z^+$, a $d$-distribution property over $[k]$ that is Lipschitz and additively separable can be learned to an accuracy of $\varepsilon$ using a sample of size $\mathcal{O}(k/(\varepsilon^3\sqrt{\log k}))$, with high probability. Our second result addresses a closely related problem– tolerant independence testing: One receives samples from the unknown joint and marginal distributions, and attempts to infer the $\ell_1$ distance between the joint distribution and the product distribution of the marginals. We show that this testing problem also admits a sample complexity sub-linear in the alphabet sizes, demonstrating the broad applicability of our approach.'
volume: 125
URL: http://proceedings.mlr.press/v125/hao20a.html
PDF: http://proceedings.mlr.press/v125/hao20a/hao20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-hao20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Hao
given: Yi
- family: Li
given: Ping
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 1817-1876
id: hao20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 1817
lastpage: 1876
published: 2020-07-15 00:00:00 +0000
- title: 'Faster Projection-free Online Learning'
abstract: ' In many online learning problems the computational bottleneck for gradient-based methods is the projection operation. For this reason, in many problems the most efficient algorithms are based on the Frank-Wolfe method, which replaces projections by linear optimization. In the general case, however, online projection-free methods require more iterations than projection-based methods: the best known regret bound scales as $T^{3/4}$. Despite significant work on various variants of the Frank-Wolfe method, this bound has remained unchanged for a decade. In this paper we give an efficient projection-free algorithm that guarantees $T^{2/3}$ regret for general online convex optimization with smooth cost functions and one linear optimization computation per iteration. As opposed to previous Frank-Wolfe approaches, our algorithm is derived using the Follow-the-Perturbed-Leader method and is analyzed using an online primal-dual framework.'
volume: 125
URL: http://proceedings.mlr.press/v125/hazan20a.html
PDF: http://proceedings.mlr.press/v125/hazan20a/hazan20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-hazan20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Hazan
given: Elad
- family: Minasyan
given: Edgar
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 1877-1893
id: hazan20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 1877
lastpage: 1893
published: 2020-07-15 00:00:00 +0000
- title: 'Near-Optimal Methods for Minimizing Star-Convex Functions and Beyond'
abstract: ' In this paper, we provide near-optimal accelerated first-order methods for minimizing a broad class of smooth nonconvex functions that are unimodal on all lines through a minimizer. This function class, which we call the class of smooth quasar-convex functions, is parameterized by a constant $$\gamma \in (0,1]$$: $$\gamma = 1$$ encompasses the classes of smooth convex and star-convex functions, and smaller values of $$\gamma$$ indicate that the function can be "more nonconvex." We develop a variant of accelerated gradient descent that computes an $$\epsilon$$-approximate minimizer of a smooth $$\gamma$$-quasar-convex function with at most $$O(\gamma^{-1} \epsilon^{-1/2} \log(\gamma^{-1} \epsilon^{-1}))$$ total function and gradient evaluations. We also derive a lower bound of $$\Omega(\gamma^{-1} \epsilon^{-1/2})$$ on the worst-case number of gradient evaluations required by any deterministic first-order method, showing that, up to a logarithmic factor, no deterministic first-order method can improve upon ours.'
volume: 125
URL: http://proceedings.mlr.press/v125/hinder20a.html
PDF: http://proceedings.mlr.press/v125/hinder20a/hinder20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-hinder20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Hinder
given: Oliver
- family: Sidford
given: Aaron
- family: Sohoni
given: Nimit
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 1894-1938
id: hinder20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 1894
lastpage: 1938
published: 2020-07-15 00:00:00 +0000
- title: 'A Greedy Anytime Algorithm for Sparse PCA'
abstract: ' The taxing computational effort that is involved in solving some high-dimensional statistical problems, in particular problems involving non-convex optimization, has popularized the development and analysis of algorithms that run efficiently (polynomial-time) but with no general guarantee on statistical consistency. In light of the ever-increasing compute power and decreasing costs, a more useful characterization of algorithms is by their ability to calibrate the invested computational effort with various characteristics of the input at hand and with the available computational resources. We propose a new greedy algorithm for the $\ell_0$-sparse PCA problem which supports the calibration principle. We provide both a rigorous analysis of our algorithm in the spiked covariance model, as well as simulation results and comparison with other existing methods. Our findings show that our algorithm recovers the spike in SNR regimes where all polynomial-time algorithms fail while running in a reasonable parallel-time on a cluster.'
volume: 125
URL: http://proceedings.mlr.press/v125/holtzman20a.html
PDF: http://proceedings.mlr.press/v125/holtzman20a/holtzman20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-holtzman20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Holtzman
given: Guy
- family: Soffer
given: Adam
- family: Vilenchik
given: Dan
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 1939-1956
id: holtzman20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 1939
lastpage: 1956
published: 2020-07-15 00:00:00 +0000
- title: 'Noise-tolerant, Reliable Active Classification with Comparison Queries'
abstract: ' With the explosion of massive, widely available unlabeled data in the past years, finding label and time efficient, robust learning algorithms has become ever more important in theory and in practice. We study the paradigm of active learning, in which algorithms with access to large pools of data may adaptively choose what samples to label in the hope of exponentially increasing efficiency. By introducing comparisons, an additional type of query comparing two points, we provide the first time and query efficient algorithms for learning non-homogeneous linear separators robust to bounded (Massart) noise. We further provide algorithms for a generalization of the popular Tsybakov low noise condition, and show how comparisons provide a strong reliability guarantee that is often impractical or impossible with only labels - returning a classifier that makes no errors with high probability.'
volume: 125
URL: http://proceedings.mlr.press/v125/hopkins20a.html
PDF: http://proceedings.mlr.press/v125/hopkins20a/hopkins20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-hopkins20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Hopkins
given: Max
- family: Kane
given: Daniel
- family: Lovett
given: Shachar
- family: Mahajan
given: Gaurav
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 1957-2006
id: hopkins20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 1957
lastpage: 2006
published: 2020-07-15 00:00:00 +0000
- title: 'Smooth Contextual Bandits: Bridging the Parametric and Non-differentiable Regret Regimes'
abstract: ' We study a nonparametric contextual bandit problem where the expected reward functions belong to a Hölder class with smoothness parameter $\beta$. We show how this interpolates between two extremes that were previously studied in isolation: non-differentiable bandits ($\beta\leq1$), where rate-optimal regret is achieved by running separate non-contextual bandits in different context regions, and parametric-response bandits (satisfying $\beta=\infty$), where rate-optimal regret can be achieved with minimal or no exploration due to infinite extrapolatability. We develop a novel algorithm that carefully adjusts to all smoothness settings and we prove its regret is rate-optimal by establishing matching upper and lower bounds, recovering the existing results at the two extremes. In this sense, our work bridges the gap between the existing literature on parametric and non-differentiable contextual bandit problems and between bandit algorithms that exclusively use global or local information, shedding light on the crucial interplay of complexity and regret in contextual bandits.'
volume: 125
URL: http://proceedings.mlr.press/v125/hu20a.html
PDF: http://proceedings.mlr.press/v125/hu20a/hu20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-hu20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Hu
given: Yichun
- family: Kallus
given: Nathan
- family: Mao
given: Xiaojie
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 2007-2010
id: hu20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 2007
lastpage: 2010
published: 2020-07-15 00:00:00 +0000
- title: 'Extrapolating the profile of a finite population'
abstract: ' We study a prototypical problem in empirical Bayes. Namely, consider a population consisting of $k$ individuals each belonging to one of $k$ types (some types can be empty). Without any structural restrictions, it is impossible to learn the composition of the full population having observed only a small (random) subsample of size $m = o(k)$. Nevertheless, we show that in the sublinear regime of $m =\omega(k/\log k)$, it is possible to consistently estimate in total variation the \emph{profile} of the population, defined as the empirical distribution of the sizes of each type, which determines many symmetric properties of the population. We also prove that in the linear regime of $m=c k$ for any constant $c$ the optimal rate is $\Theta(1/\log k)$. Our estimator is based on Wolfowitz’s minimum distance method, which entails solving a linear program (LP) of size $k$. We show that there is a single infinite-dimensional LP whose value simultaneously characterizes the risk of the minimum distance estimator and certifies its minimax optimality. The sharp convergence rate is obtained by evaluating this LP using complex-analytic techniques. '
volume: 125
URL: http://proceedings.mlr.press/v125/jana20a.html
PDF: http://proceedings.mlr.press/v125/jana20a/jana20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-jana20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Jana
given: Soham
- family: Polyanskiy
given: Yury
- family: Wu
given: Yihong
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 2011-2033
id: jana20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 2011
lastpage: 2033
published: 2020-07-15 00:00:00 +0000
- title: 'Precise Tradeoffs in Adversarial Training for Linear Regression'
abstract: ' Despite breakthrough performance, modern learning models are known to be highly vulnerable to small adversarial perturbations in their inputs. While a wide variety of recent \emph{adversarial training} methods have been effective at improving robustness to perturbed inputs (robust accuracy), often this benefit is accompanied by a decrease in accuracy on benign inputs (standard accuracy), leading to a tradeoff between often competing objectives. Complicating matters further, recent empirical evidence suggest that a variety of other factors (size and quality of training data, model size, etc.) affect this tradeoff in somewhat surprising ways. In this paper we provide a precise and comprehensive understanding of the role of adversarial training in the context of linear regression with Gaussian features. In particular, we characterize the fundamental tradeoff between the accuracies achievable by any algorithm regardless of computational power or size of the training data. Furthermore, we precisely characterize the standard/robust accuracy and the corresponding tradeoff achieved by a contemporary mini-max adversarial training approach in a high-dimensional regime where the number of data points and the parameters of the model grow in proportion to each other. Our theory for adversarial training algorithms also facilitates the rigorous study of how a variety of factors (size and quality of training data, model overparametrization etc.) affect the tradeoff between these two competing accuracies. '
volume: 125
URL: http://proceedings.mlr.press/v125/javanmard20a.html
PDF: http://proceedings.mlr.press/v125/javanmard20a/javanmard20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-javanmard20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Javanmard
given: Adel
- family: Soltanolkotabi
given: Mahdi
- family: Hassani
given: Hamed
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 2034-2078
id: javanmard20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 2034
lastpage: 2078
published: 2020-07-15 00:00:00 +0000
- title: 'Robust causal inference under covariate shift via worst-case subpopulation treatment effects'
abstract: ' We propose a notion of worst-case treatment effect (WTE) across all subpopulations of a given size, a conservative notion of topline treatment effect. Compared to the average treatment effect (ATE) that solely relies on the covariate distribution of collected data, WTE is robust to unanticipated covariate shifts, and ensures reliable inference uniformly over underrepresented minority groups. We develop a semiparametrically efficient estimator for the WTE, leveraging machine learning-based estimates of heterogenous treatment effects and propensity scores. By virtue of satisfying a key (Neyman) orthogonality property, our estimator enjoys central limit behavior—oracle rates with true nuisance parameters—even when estimates of nuisance parameters converge at slower-than-parameteric rates. In particular, this allows using black-box machine learning methods to construct asymptotically exact confidence intervals for the WTE. For both observational and randomized studies, we prove that our estimator achieves the \emph{optimal} asymptotic variance, by establishing a semiparametric efficiency lower bound. On real datasets, we illustrate the non-robustness of ATE under even small amounts distributional shift, and demonstrate that WTE allows us to guard against brittle findings that are invalidated by unanticipated covariate shifts. '
volume: 125
URL: http://proceedings.mlr.press/v125/jeong20a.html
PDF: http://proceedings.mlr.press/v125/jeong20a/jeong20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-jeong20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Jeong
given: Sookyo
- family: Namkoong
given: Hongseok
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 2079-2084
id: jeong20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 2079
lastpage: 2084
published: 2020-07-15 00:00:00 +0000
- title: 'Efficient improper learning for online logistic regression'
abstract: ' We consider the setting of online logistic regression and consider the regret with respect to the $\ell_2$-ball of radius $B$. It is known (see Hazan et al. (2014)) that any proper algorithm which has logarithmic regret in the number of samples (denoted $n$) necessarily suffers an exponential multiplicative constant in $B$. In this work, we design an efficient improper algorithm that avoids this exponential constant while preserving a logarithmic regret. Indeed, Foster et al. (2018) showed that the lower bound does not apply to improper algorithms and proposed a strategy based on exponential weights with prohibitive computational complexity. Our new algorithm based on regularized empirical risk minimization with surrogate losses satisfies a regret scaling as $O(B\log(Bn))$ with a per-round time-complexity of order $O(d^2 + \log(n))$.'
volume: 125
URL: http://proceedings.mlr.press/v125/jezequel20a.html
PDF: http://proceedings.mlr.press/v125/jezequel20a/jezequel20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-jezequel20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Jézéquel
given: Rémi
- family: Gaillard
given: Pierre
- family: Rudi
given: Alessandro
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 2085-2108
id: jezequel20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 2085
lastpage: 2108
published: 2020-07-15 00:00:00 +0000
- title: 'Gradient descent follows the regularization path for general losses'
abstract: ' Recent work across many machine learning disciplines has highlighted that standard descent methods, even without explicit regularization, do not merely minimize the training error, but also exhibit an \emph{implicit bias}. This bias is typically towards a certain regularized solution, and relies upon the details of the learning process, for instance the use of the cross-entropy loss. In this work, we show that for empirical risk minimization over linear predictors with \emph{arbitrary} convex, strictly decreasing losses, if the risk does not attain its infimum, then the gradient-descent path and the \emph{algorithm-independent} regularization path converge to the same direction (whenever either converges to a direction). Using this result, we provide a justification for the widely-used exponentially-tailed losses (such as the exponential loss or the logistic loss): while this convergence to a direction for exponentially-tailed losses is necessarily to the maximum-margin direction, other losses such as polynomially-tailed losses may induce convergence to a direction with a poor margin. '
volume: 125
URL: http://proceedings.mlr.press/v125/ji20a.html
PDF: http://proceedings.mlr.press/v125/ji20a/ji20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-ji20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Ji
given: Ziwei
- family: Dudík
given: Miroslav
- family: Schapire
given: Robert E.
- family: Telgarsky
given: Matus
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 2109-2136
id: ji20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 2109
lastpage: 2136
published: 2020-07-15 00:00:00 +0000
- title: 'Provably efficient reinforcement learning with linear function approximation'
abstract: ' Modern Reinforcement Learning (RL) is commonly applied to practical problems with an enormous number of states, where \emph{function approximation} must be deployed to approximate either the value function or the policy. The introduction of function approximation raises a fundamental set of challenges involving computational and statistical efficiency, especially given the need to manage the exploration/exploitation tradeoff. As a result, a core RL question remains open: how can we design provably efficient RL algorithms that incorporate function approximation? This question persists even in a basic setting with linear dynamics and linear rewards, for which only linear function approximation is needed. This paper presents the first provable RL algorithm with both polynomial runtime and polynomial sample complexity in this linear setting, without requiring a “simulator” or additional assumptions. Concretely, we prove that an optimistic modification of Least-Squares Value Iteration (LSVI)—a classical algorithm frequently studied in the linear setting—achieves $\tilde{\mathcal{O}}(\sqrt{d^3H^3T})$ regret, where $d$ is the ambient dimension of feature space, $H$ is the length of each episode, and $T$ is the total number of steps. Importantly, such regret is independent of the number of states and actions. '
volume: 125
URL: http://proceedings.mlr.press/v125/jin20a.html
PDF: http://proceedings.mlr.press/v125/jin20a/jin20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-jin20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Jin
given: Chi
- family: Yang
given: Zhuoran
- family: Wang
given: Zhaoran
- family: Jordan
given: Michael I
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 2137-2143
id: jin20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 2137
lastpage: 2143
published: 2020-07-15 00:00:00 +0000
- title: 'Finite Time Analysis of Linear Two-timescale Stochastic Approximation with Markovian Noise'
abstract: ' Linear two-timescale stochastic approximation (SA) scheme is an important class of algorithms which has become popular in reinforcement learning (RL), particularly for the policy evaluation problem. Recently, a number of works have been devoted to establishing the finite time analysis of the scheme, especially under the Markovian (non-i.i.d.) noise settings that are ubiquitous in practice. In this paper, we provide a finite-time analysis for linear two timescale SA. Our bounds show that there is no discrepancy in the convergence rate between Markovian and martingale noise, only the constants are affected by the mixing time of the Markov chain. With an appropriate step size schedule, the transient term in the expected error bound is $o(1/k^c)$ and the steady-state term is ${\cal O}(1/k)$, where $c>1$ and $k$ is the iteration number. Furthermore, we present an asymptotic expansion of the expected error with a matching lower bound of $\Omega(1/k)$. A simple numerical experiment is presented to support our theory.'
volume: 125
URL: http://proceedings.mlr.press/v125/kaledin20a.html
PDF: http://proceedings.mlr.press/v125/kaledin20a/kaledin20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-kaledin20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Kaledin
given: Maxim
- family: Moulines
given: Eric
- family: Naumov
given: Alexey
- family: Tadic
given: Vladislav
- family: Wai
given: Hoi-To
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 2144-2203
id: kaledin20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 2144
lastpage: 2203
published: 2020-07-15 00:00:00 +0000
- title: 'Private Mean Estimation of Heavy-Tailed Distributions'
abstract: ' We give new upper and lower bounds on the minimax sample complexity of differentially private mean estimation of distributions with bounded $k$-th moments. Roughly speaking, in the univariate case, we show that $$n = \Theta\left(\frac{1}{\alpha^2} + \frac{1}{\alpha^{\frac{k}{k-1}}\varepsilon}\right)$$ samples are necessary and sufficient to estimate the mean to $\alpha$-accuracy under $\varepsilon$-differential privacy, or any of its common relaxations. This result demonstrates a qualitatively different behavior compared to estimation absent privacy constraints, for which the sample complexity is identical for all $k \geq 2$. We also give algorithms for the multivariate setting whose sample complexity is a factor of $O(d)$ larger than the univariate case.'
volume: 125
URL: http://proceedings.mlr.press/v125/kamath20a.html
PDF: http://proceedings.mlr.press/v125/kamath20a/kamath20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-kamath20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Kamath
given: Gautam
- family: Singhal
given: Vikrant
- family: Ullman
given: Jonathan
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 2204-2235
id: kamath20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 2204
lastpage: 2235
published: 2020-07-15 00:00:00 +0000
- title: 'Approximate is Good Enough: Probabilistic Variants of Dimensional and Margin Complexity'
abstract: ' We present and study approximate notions of dimensional and margin complexity, which correspond to the minimal dimension or norm of an embedding required to {\em approximate}, rather then exactly represent, a given hypothesis class. We show that such notions are not only sufficient for learning using linear predictors or a kernel, but unlike the exact variants, are also necessary. Thus they are better suited for discussing limitations of linear or kernel methods.'
volume: 125
URL: http://proceedings.mlr.press/v125/kamath20b.html
PDF: http://proceedings.mlr.press/v125/kamath20b/kamath20b.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-kamath20b.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Kamath
given: Pritish
- family: Montasser
given: Omar
- family: Srebro
given: Nathan
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 2236-2262
id: kamath20b
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 2236
lastpage: 2262
published: 2020-07-15 00:00:00 +0000
- title: 'Privately Learning Thresholds: Closing the Exponential Gap'
abstract: ' We study the sample complexity of learning threshold functions under the constraint of differential privacy. It is assumed that each labeled example in the training data is the information of one individual and we would like to come up with a generalizing hypothesis $h$ while guaranteeing differential privacy for the individuals. Intuitively, this means that any single labeled example in the training data should not have a significant effect on the choice of the hypothesis. This problem has received much attention recently; unlike the non-private case, where the sample complexity is independent of the domain size and just depends on the desired accuracy and confidence, for private learning the sample complexity must depend on the domain size $X$ (even for approximate differential privacy). Alon et al. (STOC 2019) showed a lower bound of $\Omega(\log^*|X|)$ on the sample complexity and Bun et al. (FOCS 2015) presented an approximate-private learner with sample complexity $\tilde{O}\left(2^{\log^*|X|}\right)$. In this work we reduce this gap significantly, almost settling the sample complexity. We first present a new upper bound (algorithm) of $\tilde{O}\left(\left(\log^*|X|\right)^2\right)$ on the sample complexity and then present an improved version with sample complexity $\tilde{O}\left(\left(\log^*|X|\right)^{1.5}\right)$. Our algorithm is constructed for the related interior point problem, where the goal is to find a point between the largest and smallest input elements. It is based on selecting an input-dependent hash function and using it to embed the database into a domain whose size is reduced logarithmically; this results in a new database, an interior point of which can be used to generate an interior point of the original database in a differentially private manner.'
volume: 125
URL: http://proceedings.mlr.press/v125/kaplan20a.html
PDF: http://proceedings.mlr.press/v125/kaplan20a/kaplan20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-kaplan20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Kaplan
given: Haim
- family: Ligett
given: Katrina
- family: Mansour
given: Yishay
- family: Naor
given: Moni
- family: Stemmer
given: Uri
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 2263-2285
id: kaplan20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 2263
lastpage: 2285
published: 2020-07-15 00:00:00 +0000
- title: 'Online Learning with Vector Costs and Bandits with Knapsacks'
abstract: ' We introduce online learning with vector costs ($OLVC_p$) where in each time step $t \in \{1,\ldots, T\}$, we need to play an action $i \in \{1,\ldots,n\}$ that incurs an unknown vector cost in $[0,1]^d$. The goal of the online algorithm is to minimize the $\ell_p$ norm of the sum of its cost vectors. This captures the classical online learning setting for $d=1$, and is interesting for general $d$ because of applications like online scheduling where we want to balance the load between different machines (dimensions). We study $OLVC_p$ in both stochastic and adversarial arrival settings, and give a general procedure to reduce the problem from $d$ dimensions to a single dimension. This allows us to use classical online learning algorithms in both full and bandit feedback models to obtain (near) optimal results. In particular, we obtain a single algorithm (up to the choice of learning rate) that gives sublinear regret for stochastic arrivals and a tight $O(\min\{p, \log d\})$ competitive ratio for adversarial arrivals. The $OLVC_p$ problem also occurs as a natural subproblem when trying to solve the popular Bandits with Knapsacks (BWK) problem. This connection allows us to use our $OLVC_p$ techniques to obtain (near) optimal results for BWK in both stochastic and adversarial settings. In particular, we obtain a tight $O(\log d \cdot \log T)$ competitive ratio algorithm for adversarial BWK, which improves over the $O(d \cdot \log T)$ competitive ratio algorithm of Immorlica et al. (2019).'
volume: 125
URL: http://proceedings.mlr.press/v125/kesselheim20a.html
PDF: http://proceedings.mlr.press/v125/kesselheim20a/kesselheim20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-kesselheim20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Kesselheim
given: Thomas
- family: Singla
given: Sahil
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 2286-2305
id: kesselheim20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 2286
lastpage: 2305
published: 2020-07-15 00:00:00 +0000
- title: 'Universal Approximation with Deep Narrow Networks'
abstract: ' The classical Universal Approximation Theorem holds for neural networks of arbitrary width and bounded depth. Here we consider the natural ‘dual’ scenario for networks of bounded width and arbitrary depth. Precisely, let $n$ be the number of inputs neurons, $m$ be the number of output neurons, and let $\rho$ be any nonaffine continuous function, with a continuous nonzero derivative at some point. Then we show that the class of neural networks of arbitrary depth, width $n + m + 2$, and activation function $\rho$, is dense in $C(K; \mathbb{R}^m)$ for $K \subseteq \mathbb{R}^n$ with $K$ compact. This covers every activation function possible to use in practice, and also includes polynomial activation functions, which is unlike the classical version of the theorem, and provides a qualitative difference between deep narrow networks and shallow wide networks. We then consider several extensions of this result. In particular we consider nowhere differentiable activation functions, density in noncompact domains with respect to the $L^p$-norm, and how the width may be reduced to just $n + m + 1$ for ‘most’ activation functions.'
volume: 125
URL: http://proceedings.mlr.press/v125/kidger20a.html
PDF: http://proceedings.mlr.press/v125/kidger20a/kidger20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-kidger20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Kidger
given: Patrick
- family: Lyons
given: Terry
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 2306-2327
id: kidger20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 2306
lastpage: 2327
published: 2020-07-15 00:00:00 +0000
- title: 'Information Directed Sampling for Linear Partial Monitoring'
abstract: ' Partial monitoring is a rich framework for sequential decision making under uncertainty that generalizes many well known bandit models, including linear, combinatorial and dueling bandits. We introduce {\em information directed sampling} (IDS) for stochastic partial monitoring with a linear reward and observation structure. IDS achieves adaptive worst-case regret rates that depend on precise observability conditions of the game. Moreover, we prove lower bounds that classify the minimax regret of all finite games into four possible regimes. IDS achieves the optimal rate in all cases up to logarithmic factors, without tuning any hyper-parameters. We further extend our results to the contextual and the kernelized setting, which significantly increases the range of possible applications.'
volume: 125
URL: http://proceedings.mlr.press/v125/kirschner20a.html
PDF: http://proceedings.mlr.press/v125/kirschner20a/kirschner20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-kirschner20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Kirschner
given: Johannes
- family: Lattimore
given: Tor
- family: Krause
given: Andreas
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 2328-2369
id: kirschner20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 2328
lastpage: 2369
published: 2020-07-15 00:00:00 +0000
- title: 'New Potential-Based Bounds for Prediction with Expert Advice'
abstract: ' This work addresses the classic machine learning problem of online prediction with expert advice. We consider the finite-horizon version of this zero-sum, two-person game. Using verification arguments from optimal control theory, we view the task of finding better lower and upper bounds on the value of the game (regret) as the problem of finding better sub- and supersolutions of certain partial differential equations (PDEs). These sub- and supersolutions serve as the potentials for player and adversary strategies, which lead to the corresponding bounds. To get explicit bounds, we use closed-form solutions of specific PDEs. Our bounds hold for any given number of experts and horizon; in certain regimes (which we identify) they improve upon the previous state of the art. For two and three experts, our bounds provide the optimal leading order term. '
volume: 125
URL: http://proceedings.mlr.press/v125/kobzar20a.html
PDF: http://proceedings.mlr.press/v125/kobzar20a/kobzar20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-kobzar20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Kobzar
given: Vladimir A.
- family: Kohn
given: Robert V.
- family: Wang
given: Zhilei
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 2370-2405
id: kobzar20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 2370
lastpage: 2405
published: 2020-07-15 00:00:00 +0000
- title: 'On Suboptimality of Least Squares with Application to Estimation of Convex Bodies'
abstract: ' We develop a technique for establishing lower bounds on the sample complexity of Least Squares (or, Empirical Risk Minimization) for large classes of functions. As an application, we settle an open problem regarding optimality of Least Squares in estimating a convex set from noisy support function measurements in dimension $d\geq 6$. Specifically, we establish that Least Squares is mimimax sub-optimal, and achieves a rate of $\tilde{\Theta}_d(n^{-2/(d-1)})$ whereas the minimax rate is $\Theta_d(n^{-4/(d+3)})$.'
volume: 125
URL: http://proceedings.mlr.press/v125/kur20a.html
PDF: http://proceedings.mlr.press/v125/kur20a/kur20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-kur20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Kur
given: Gil
- family: Rakhlin
given: Alexander
- family: Guntuboyina
given: Adityanand
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 2406-2424
id: kur20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 2406
lastpage: 2424
published: 2020-07-15 00:00:00 +0000
- title: 'The EM Algorithm gives Sample-Optimality for Learning Mixtures of Well-Separated Gaussians'
abstract: ' We consider the problem of spherical Gaussian Mixture models with $k \geq 3$ components when the components are well separated. A fundamental previous result established that separation of $\Omega(\sqrt{\log k})$ is necessary and sufficient for identifiability of the parameters with \textit{polynomial} sample complexity (Regev and Vijayaraghavan, 2017). In the same context, we show that $\tilde{O} (kd/\epsilon^2)$ samples suffice for any $\epsilon \lesssim 1/k$, closing the gap from polynomial to linear, and thus giving the first optimal sample upper bound for the parameter estimation of well-separated Gaussian mixtures. We accomplish this by proving a new result for the Expectation-Maximization (EM) algorithm: we show that EM converges locally, under separation $\Omega(\sqrt{\log k})$. The previous best-known guarantee required $\Omega(\sqrt{k})$ separation (Yan, et al., 2017). Unlike prior work, our results do not assume or use prior knowledge of the (potentially different) mixing weights or variances of the Gaussian components. Furthermore, our results show that the finite-sample error of EM does not depend on non-universal quantities such as pairwise distances between means of Gaussian components.'
volume: 125
URL: http://proceedings.mlr.press/v125/kwon20a.html
PDF: http://proceedings.mlr.press/v125/kwon20a/kwon20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-kwon20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Kwon
given: Jeongyeol
- family: Caramanis
given: Constantine
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 2425-2487
id: kwon20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 2425
lastpage: 2487
published: 2020-07-15 00:00:00 +0000
- title: 'Exploration by Optimisation in Partial Monitoring'
abstract: ' We provide a novel algorithm for adversarial k-action d-outcome partial monitoring that is adaptive, intuitive and efficient. The highlight is that for the non-degenerate locally observable games, the n-round minimax regret is bounded by 6m k^(3/2) sqrt(n log(k)), where m is the number of signals. This matches the best known information-theoretic upper bound derived via Bayesian minimax duality. The same algorithm also achieves near-optimal regret for full information, bandit and globally observable games. High probability bounds and simple experiments are also provided.'
volume: 125
URL: http://proceedings.mlr.press/v125/lattimore20a.html
PDF: http://proceedings.mlr.press/v125/lattimore20a/lattimore20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-lattimore20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Lattimore
given: Tor
- family: Szepesvári
given: Csaba
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 2488-2515
id: lattimore20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 2488
lastpage: 2515
published: 2020-07-15 00:00:00 +0000
- title: 'A Closer Look at Small-loss Bounds for Bandits with Graph Feedback'
abstract: ' We study {\it small-loss} bounds for adversarial multi-armed bandits with graph feedback, that is, adaptive regret bounds that depend on the loss of the best arm or related quantities, instead of the total number of rounds. We derive the first small-loss bound for general strongly observable graphs, resolving an open problem of Lykouris et al. (2018). Specifically, we develop an algorithm with regret $\mathcal{\tilde{O}}(\sqrt{\kappa L_*})$ where $\kappa$ is the clique partition number and $L_*$ is the loss of the best arm, and for the special case of self-aware graphs where every arm has a self-loop, we improve the regret to $\mathcal{\tilde{O}}(\min\{\sqrt{\alpha T}, \sqrt{\kappa L_*}\})$ where $\alpha \leq \kappa$ is the independence number. Our results significantly improve and extend those by Lykouris et al. (2018) who only consider self-aware undirected graphs. Furthermore, we also take the first attempt at deriving small-loss bounds for weakly observable graphs. We first prove that no typical small-loss bounds are achievable in this case, and then propose algorithms with alternative small-loss bounds in terms of the loss of some specific subset of arms. A surprising side result is that $\mathcal{\tilde{O}}(\sqrt{T})$ regret is achievable even for weakly observable graphs as long as the best arm has a self-loop. Our algorithms are based on the Online Mirror Descent framework but require a suite of novel techniques that might be of independent interest. Moreover, all our algorithms can be made parameter-free without the knowledge of the environment.'
volume: 125
URL: http://proceedings.mlr.press/v125/lee20a.html
PDF: http://proceedings.mlr.press/v125/lee20a/lee20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-lee20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Lee
given: Chung-Wei
- family: Luo
given: Haipeng
- family: Zhang
given: Mengxiao
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 2516-2564
id: lee20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 2516
lastpage: 2564
published: 2020-07-15 00:00:00 +0000
- title: 'Logsmooth Gradient Concentration and Tighter Runtimes for Metropolized Hamiltonian Monte Carlo'
abstract: ' We show that the gradient norm $\norm{\nabla f(x)}$ for $x \sim \exp(-f(x))$, where $f$ is strongly convex and smooth, concentrates tightly around its mean. This removes a barrier in the prior state-of-the-art analysis for the well-studied Metropolized Hamiltonian Monte Carlo (HMC) algorithm for sampling from a strongly logconcave distribution. We correspondingly demonstrate that Metropolized HMC mixes in $\tilde{O}(\kappa d)$ iterations, improving upon the $\tilde{O}(\kappa^{1.5}\sqrt{d} + \kappa d)$ runtime of prior work by a factor $(\kappa/d)^{1/2}$ when the condition number $\kappa$ is large. Our mixing time analysis introduces several techniques which to our knowledge have not appeared in the literature and may be of independent interest, including restrictions to a nonconvex set with good conductance behavior, and a new reduction technique for boosting a constant-accuracy total variation guarantee under weak warmness assumptions. This is the first high-accuracy mixing time result for logconcave distributions using only first-order function information which achieves linear dependence on $\kappa$; we also give evidence that this dependence is likely to be necessary for standard Metropolized first-order methods.'
volume: 125
URL: http://proceedings.mlr.press/v125/lee20b.html
PDF: http://proceedings.mlr.press/v125/lee20b/lee20b.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-lee20b.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Lee
given: Yin Tat
- family: Shen
given: Ruoqi
- family: Tian
given: Kevin
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 2565-2597
id: lee20b
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 2565
lastpage: 2597
published: 2020-07-15 00:00:00 +0000
- title: 'A Fast Spectral Algorithm for Mean Estimation with Sub-Gaussian Rates'
abstract: ' We study the algorithmic problem of estimating the mean of a heavy-tailed random vector in R^d, given n i.i.d. samples. The goal is to design an efficient estimator that attains the optimal sub-gaussian error bound, only assuming that the random vector has bounded mean and covariance. Polynomial-time solutions to this problem are known but have high runtime due to their use of semi-definite programming (SDP). Moreover, conceptually, it remains open whether convex relaxation is truly necessary for this problem. In this work, we show that it is possible to go beyond SDP and achieve better computational efficiency. In particular, we provide a spectral algorithm that achieves the optimal statistical performance and runs in time O ( n^2 d ), improving upon the previous fastest runtime O( n^{3.5}+ n^2 d ) by Cherapanamjeri et.al. (COLT ’19). Our algorithm is spectral in that it only requires (approximate) eigenvector computations, which can be implemented very efficiently by, for example, power iteration or the Lanczos method. At the core of our algorithm is a novel connection between the furthest hyperplane problem introduced by Karnin et. al. (COLT ’12) and a structural lemma on heavy-tailed distributions by Lugosi and Mendelson (Ann. Stat. ’19). This allows us to iteratively reduce the estimation error at a geometric rate using only the information derived from the top singular vector of the data matrix, leading to a significantly faster running time.'
volume: 125
URL: http://proceedings.mlr.press/v125/lei20a.html
PDF: http://proceedings.mlr.press/v125/lei20a/lei20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-lei20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Lei
given: Zhixian
- family: Luh
given: Kyle
- family: Venkat
given: Prayaag
- family: Zhang
given: Fred
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 2598-2612
id: lei20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 2598
lastpage: 2612
published: 2020-07-15 00:00:00 +0000
- title: 'Learning Over-Parametrized Two-Layer Neural Networks beyond NTK'
abstract: ' We consider the dynamic of gradient descent for learning a two-layer neural network. We assume the input $x\in\mathbb{R}^d$ is drawn from a Gaussian distribution and the label of $x$ satisfies $f^{\star}(x) = a^{\top}|W^{\star}x|$, where $a\in\mathbb{R}^d$ is a nonnegative vector and $W^{\star} \in\mathbb{R}^{d\times d}$ is an orthonormal matrix. We show that an \emph{over-parameterized} two layer neural network with ReLU activation, trained by gradient descent from \emph{random initialization}, can provably learn the ground truth network with population loss at most $o(1/d)$ in polynomial time with polynomial samples. On the other hand, we prove that any kernel method, including Neural Tangent Kernel, with a polynomial number of samples in $d$, has population loss at least $\Omega(1 / d)$.'
volume: 125
URL: http://proceedings.mlr.press/v125/li20a.html
PDF: http://proceedings.mlr.press/v125/li20a/li20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-li20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Li
given: Yuanzhi
- family: Ma
given: Tengyu
- family: Zhang
given: Hongyang R.
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 2613-2682
id: li20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 2613
lastpage: 2682
published: 2020-07-15 00:00:00 +0000
- title: 'On the Multiple Descent of Minimum-Norm Interpolants and Restricted Lower Isometry of Kernels'
abstract: ' We study the risk of minimum-norm interpolants of data in Reproducing Kernel Hilbert Spaces. Our upper bounds on the risk are of a multiple-descent shape for the various scalings of $d = n^{\alpha}$, $\alpha\in(0,1)$, for the input dimension $d$ and sample size $n$. Empirical evidence supports our finding that minimum-norm interpolants in RKHS can exhibit this unusual non-monotonicity in sample size; furthermore, locations of the peaks in our experiments match our theoretical predictions. Since gradient flow on appropriately initialized wide neural networks converges to a minimum-norm interpolant with respect to a certain kernel, our analysis also yields novel estimation and generalization guarantees for these over-parametrized models. At the heart of our analysis is a study of spectral properties of the random kernel matrix restricted to a filtration of eigen-spaces of the population covariance operator, and may be of independent interest. '
volume: 125
URL: http://proceedings.mlr.press/v125/liang20a.html
PDF: http://proceedings.mlr.press/v125/liang20a/liang20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-liang20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Liang
given: Tengyuan
- family: Rakhlin
given: Alexander
- family: Zhai
given: Xiyu
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 2683-2711
id: liang20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 2683
lastpage: 2711
published: 2020-07-15 00:00:00 +0000
- title: 'Learning Entangled Single-Sample Gaussians in the Subset-of-Signals Model'
abstract: ' In the setting of entangled single-sample distributions, the goal is to estimate some common parameter shared by a family of $n$ distributions, given one single sample from each distribution. This paper studies mean estimation for entangled single-sample Gaussians that have a common mean but different unknown variances. We propose the subset-of-signals model where an unknown subset of $m$ variances are bounded by 1 while there are no assumptions on the other variances. In this model, we analyze a simple and natural method based on iteratively averaging the truncated samples, and show that the method achieves error $O \left(\frac{\sqrt{n\ln n}}{m}\right)$ with high probability when $m=\Omega(\sqrt{n\ln n})$, slightly improving existing bounds for this range of $m$. We further prove lower bounds, showing that the error is $\Omega\left(\left(\frac{n}{m^4}\right)^{1/2}\right)$ when $m$ is between $\Omega(\ln n)$ and $O(n^{1/4})$, and the error is $\Omega\left(\left(\frac{n}{m^4}\right)^{1/6}\right)$ when $m$ is between $\Omega(n^{1/4})$ and $O(n^{1 - \epsilon})$ for an arbitrarily small $\epsilon>0$, improving existing lower bounds and extending to a wider range of $m$.'
volume: 125
URL: http://proceedings.mlr.press/v125/liang20b.html
PDF: http://proceedings.mlr.press/v125/liang20b/liang20b.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-liang20b.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Liang
given: Yingyu
- family: Yuan
given: Hui
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 2712-2737
id: liang20b
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 2712
lastpage: 2737
published: 2020-07-15 00:00:00 +0000
- title: 'Near-Optimal Algorithms for Minimax Optimization'
abstract: ' This paper resolves a longstanding open question pertaining to the design of near-optimal first-order algorithms for smooth and strongly-convex-strongly-concave minimax problems. Current state-of-the-art first-order algorithms find an approximate Nash equilibrium using $\tilde{O}(\kappa_x+\kappa_y)$ or $\tilde{O}(\text{min}\{\kappa_x\sqrt{\kappa_y}, \sqrt{\kappa_x}\kappa_y\})$ gradient evaluations, where $\kappa_x$ and $\kappa_y$ are the condition numbers for the strong-convexity and strong-concavity assumptions. A gap still remains between these results and the best existing lower bound $\tilde{\Omega}(\sqrt{\kappa_x\kappa_y})$. This paper presents the first algorithm with $\tilde{O}(\sqrt{\kappa_x\kappa_y})$ gradient complexity, matching the lower bound up to logarithmic factors. Our algorithm is designed based on an accelerated proximal point method and an accelerated solver for minimax proximal steps. It can be easily extended to the settings of strongly-convex-concave, convex-concave, nonconvex-strongly-concave, and nonconvex-concave functions. This paper also presents algorithms that match or outperform all existing methods in these settings in terms of gradient complexity, up to logarithmic factors.'
volume: 125
URL: http://proceedings.mlr.press/v125/lin20a.html
PDF: http://proceedings.mlr.press/v125/lin20a/lin20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-lin20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Lin
given: Tianyi
- family: Jin
given: Chi
- family: Jordan
given: Michael I.
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 2738-2779
id: lin20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 2738
lastpage: 2779
published: 2020-07-15 00:00:00 +0000
- title: 'Better Algorithms for Estimating Non-Parametric Models in Crowd-Sourcing and Rank Aggregation'
abstract: ' Motivated by applications in crowd-sourcing and rank aggregation, a recent line of work has studied the problem of estimating an $n \times n$ bivariate isotonic matrix with an unknown permutation acting on its rows (and possibly another unknown permutation acting on its columns) from partial and noisy observations. There are wide and persistent computational vs. statistical gaps for this problem. It is known that the minimax optimal rate is $\widetilde{O}(n^{-1})$ when error is measured in average squared Frobenius norm. However the best known polynomial time computable estimator due to \cite{coltpaper} achieves the rate $\widetilde{O}(n^{-\frac{3}{4}})$, and this is the natural barrier to approaches based on using local statistics to figure out the relative order of pairs of rows without using information from the rest of the matrix. Here we introduce a framework for exploiting global information in shape-constrained estimation problems. In the case when only the rows are permuted, we give an algorithm that achieves error rate $O(n^{-1 + o(1)})$, which essentially closes the computational vs. statistical gap for this problem. When both the rows and columns are permuted, we give an improved algorithm that achieves error rate $O(n^{-\frac{5}{6} + o(1)})$. Additionally, all of our algorithms run in nearly linear time. '
volume: 125
URL: http://proceedings.mlr.press/v125/liu20a.html
PDF: http://proceedings.mlr.press/v125/liu20a/liu20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-liu20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Liu
given: Allen
- family: Moitra
given: Ankur
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 2780-2829
id: liu20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 2780
lastpage: 2829
published: 2020-07-15 00:00:00 +0000
- title: 'Tight Lower Bounds for Combinatorial Multi-Armed Bandits'
abstract: ' The Combinatorial Multi-Armed Bandit problem is a sequential decision-making problem in which an agent selects a set of arms on each round, observes feedback for each of these arms and aims to maximize a known reward function of the arms it chose. While previous work proved regret upper bounds in this setting for general reward functions, only a few works provided matching lower bounds, all for specific reward functions. In this work, we prove regret lower bounds for combinatorial bandits that hold under mild assumptions for all smooth reward functions. We derive both problem-dependent and problem-independent bounds and show that the recently proposed Gini-weighted smoothness parameter (Merlis and Mannor, 2019) also determines the lower bounds for monotone reward functions. Notably, this implies that our lower bounds are tight up to log-factors.'
volume: 125
URL: http://proceedings.mlr.press/v125/merlis20a.html
PDF: http://proceedings.mlr.press/v125/merlis20a/merlis20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-merlis20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Merlis
given: Nadav
- family: Mannor
given: Shie
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 2830-2857
id: merlis20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 2830
lastpage: 2857
published: 2020-07-15 00:00:00 +0000
- title: 'Lipschitz and Comparator-Norm Adaptivity in Online Learning'
abstract: ' We study Online Convex Optimization in the unbounded setting where neither predictions nor gradient are constrained. The goal is to simultaneously adapt to both the sequence of gradients and the comparator. We first develop parameter-free and scale-free algorithms for a simplified setting with hints. We present two versions: the first adapts to the squared norms of both comparator and gradients separately using $O(d)$ time per round, the second adapts to their squared inner products (which measure variance only in the comparator direction) in time $O(d^3)$ per round. We then generalize two prior reductions to the unbounded setting; one to not need hints, and a second to deal with the range ratio problem (which already arises in prior work). We discuss their optimality in light of prior and new lower bounds. We apply our methods to obtain sharper regret bounds for scale-invariant online prediction with linear models. '
volume: 125
URL: http://proceedings.mlr.press/v125/mhammedi20a.html
PDF: http://proceedings.mlr.press/v125/mhammedi20a/mhammedi20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-mhammedi20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Mhammedi
given: Zakaria
- family: Koolen
given: Wouter M.
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 2858-2887
id: mhammedi20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 2858
lastpage: 2887
published: 2020-07-15 00:00:00 +0000
- title: 'Information Theoretic Optimal Learning of Gaussian Graphical Models'
abstract: ' What is the optimal number of independent observations from which a sparse Gaussian Graphical Model can be correctly recovered? Information-theoretic arguments provide a lower bound on the minimum number of samples necessary to perfectly identify the support of any multivariate normal distribution as a function of model parameters. For a model defined on a sparse graph with $p$ nodes, a maximum degree $d$ and minimum normalized edge strength $\kappa$, this necessary number of samples scales at least as $d \log p/\kappa^2$. The sample complexity requirements of existing methods for perfect graph reconstruction exhibit dependency on additional parameters that do not enter in the lower bound. The question of whether the lower bound is tight and achievable by a polynomial time algorithm remains open. In this paper, we constructively answer this question and propose an algorithm, termed DICE, whose sample complexity matches the information-theoretic lower bound up to a universal constant factor. We also propose a related algorithm SLICE that has a slightly higher sample complexity, but can be implemented as a mixed integer quadratic program which makes it attractive in practice. Importantly, SLICE retains a critical advantage of DICE in that its sample complexity only depends on quantities present in the information theoretic lower bound. We anticipate that this result will stimulate future search of computationally efficient sample-optimal algorithms.'
volume: 125
URL: http://proceedings.mlr.press/v125/misra20a.html
PDF: http://proceedings.mlr.press/v125/misra20a/misra20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-misra20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Misra
given: Sidhant
- family: Vuffray
given: Marc
- family: Lokhov
given: Andrey Y.
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 2888-2909
id: misra20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 2888
lastpage: 2909
published: 2020-07-15 00:00:00 +0000
- title: 'Parallels Between Phase Transitions and Circuit Complexity?'
abstract: ' In many natural average-case problems, there are or there are believed to be critical values in the parameter space where the structure of the space of solutions changes in a fundamental way. These phase transitions are often believed to coincide with drastic changes in the computational complexity of the associated problem. In this work, we study the circuit complexity of inference in the broadcast tree model, which has important applications in phylogenetic reconstruction and close connections to community detection. We establish a number of qualitative connections between phase transitions and circuit complexity in this model. Specifically we show that there is a $\mathbf{TC}^0$ circuit that competes with the Bayes optimal predictor in some range of parameters above the Kesten-Stigum bound. We also show that there is a $16$ label broadcast tree model beneath the Kesten-Stigum bound in which it is possible to accurately guess the label of the root, but beating random guessing is $\mathbf{NC}^1$-hard on average. The key to locating phase transitions is often to study some intrinsic notions of complexity associated with belief propagation \— e.g. where do linear statistics fail, or when is the posterior sensitive to noise? Ours is the first work to study the complexity of belief propagation in a way that is grounded in circuit complexity.'
volume: 125
URL: http://proceedings.mlr.press/v125/moitra20a.html
PDF: http://proceedings.mlr.press/v125/moitra20a/moitra20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-moitra20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Moitra
given: Ankur
- family: Mossel
given: Elchanan
- family: Sandon
given: Colin
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 2910-2946
id: moitra20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 2910
lastpage: 2946
published: 2020-07-15 00:00:00 +0000
- title: 'On Linear Stochastic Approximation: Fine-grained Polyak-Ruppert and Non-Asymptotic Concentration'
abstract: ' We undertake a precise study of the asymptotic and non-asymptotic properties of stochastic approximation procedures with Polyak-Ruppert averaging for solving a linear system $\bar{A} \theta = \bar{b}$. When the matrix $\bar{A}$ is Hurwitz, we prove a central limit theorem (CLT) for the averaged iterates with fixed step size and number of iterations going to infinity. The CLT characterizes the exact asymptotic covariance matrix, which is the sum of the classical Polyak-Ruppert covariance and a correction term that scales with the step size. Under assumptions on the tail of the noise distribution, we prove a non-asymptotic concentration inequality whose main term matches the covariance in CLT in any direction, up to universal constants. When the matrix $\bar{A}$ is not Hurwitz but only has non-negative real parts in its eigenvalues, we prove that the averaged LSA procedure actually achieves an $O(1/T)$ rate in mean-squared error. Our results provide a more refined understanding of linear stochastic approximation in both the asymptotic and non-asymptotic settings. We also show various applications of the main results, including the study of momentum-based stochastic gradient methods as well as temporal difference algorithms in reinforcement learning.'
volume: 125
URL: http://proceedings.mlr.press/v125/mou20a.html
PDF: http://proceedings.mlr.press/v125/mou20a/mou20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-mou20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Mou
given: Wenlong
- family: Li
given: Chris Junchi
- family: Wainwright
given: Martin J
- family: Bartlett
given: Peter L
- family: Jordan
given: Michael I
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 2947-2997
id: mou20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 2947
lastpage: 2997
published: 2020-07-15 00:00:00 +0000
- title: 'Extending Learnability to Auxiliary-Input Cryptographic Primitives and Meta-PAC Learning'
abstract: ' We investigate the meaning of efficient learnability from several different perspectives. The purpose is to give new insights into central problems in computational learning theory (CoLT). Specifically, we discuss the following two questions related to efficient PAC learnability. First, we investigate the gap between PAC learnability for polynomial-size circuits and weak cryptographic primitives taking auxiliary-input. Applebaum et al. observed that such a weak primitive is enough to show the hardness of PAC learning. However, the opposite direction is still unknown. In this paper, we introduce the following two notions: (1) a variant model of PAC learning whose hardness corresponds to auxiliary-input one-way functions; (2) a variant of a hitting set generator corresponding to the hardness of PAC learning. The equivalence gives a clearer insight into the gap between the hardness of learning and weak cryptographic primitives. Second, we discuss why proving efficient learnability is difficult. This question is natural because few classes are known to be polynomially learnable at present. In this paper, we formulate a task of determining efficient learnability as a meta-PAC learning problem and show that our meta-PAC learning is exactly as hard as PAC learning. Our result insists on one possibility: a hard-to-learn instance itself yields the hardness of proving efficient learnability. Our technical contribution is to give (1) a general framework for translating the hardness of PAC learning into auxiliary-input primitives, and (2) a new formulation to discuss the hardness of determining efficient learnability. Our work yields new important frontiers related to CoLT, including investigation of the learning hierarchy. '
volume: 125
URL: http://proceedings.mlr.press/v125/nanashima20a.html
PDF: http://proceedings.mlr.press/v125/nanashima20a/nanashima20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-nanashima20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Nanashima
given: Mikito
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 2998-3029
id: nanashima20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 2998
lastpage: 3029
published: 2020-07-15 00:00:00 +0000
- title: 'Fast Rates for Online Prediction with Abstention'
abstract: ' In the setting of sequential prediction of individual $(0, 1)$-sequences with expert advice, we show that by allowing the learner to abstain from the prediction by paying a cost marginally smaller than $0.5$ (say, $0.49$), it is possible to achieve expected regret bounds that are independent of the time horizon T. We exactly characterize the dependence on the abstention cost $c$ and the number of experts N by providing matching upper and lower bounds of order $\frac{log N} {1−2c}$, which is to be contrasted with the best possible rate of $\sqrt{T\log N}$ that is available without the option to abstain. We also discuss various extensions of our model, including a setting where the sequence of abstention costs can change arbitrarily over time, where we show regret bounds interpolating between the slow and the fast rates mentioned above, under some natural assumptions on the sequence of abstention costs'
volume: 125
URL: http://proceedings.mlr.press/v125/neu20a.html
PDF: http://proceedings.mlr.press/v125/neu20a/neu20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-neu20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Neu
given: Gergely
- family: Zhivotovskiy
given: Nikita
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 3030-3048
id: neu20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 3030
lastpage: 3048
published: 2020-07-15 00:00:00 +0000
- title: 'Efficient and robust algorithms for adversarial linear contextual bandits'
abstract: ' We consider an adversarial variant of the classic $K$-armed linear contextual bandit problem where the sequence of loss functions associated with each arm are allowed to change without restriction over time. Under the assumption that the $d$-dimensional contexts are generated i.i.d. at random from a known distribution, we develop computationally efficient algorithms based on the classic Exp3 algorithm. Our first algorithm, RealLinExp3, is shown to achieve a regret guarantee of $\widetilde{O}(\sqrt{KdT})$ over $T$ rounds, which matches the best known lower bound for this problem. Our second algorithm, RobustLinExp3, is shown to be robust to misspecification, in that it achieves a regret bound of $\widetilde{O}((Kd)^{1/3}T^{2/3}) + \varepsilon \sqrt{d} T$ if the true reward function is linear up to an additive nonlinear error uniformly bounded in absolute value by $\varepsilon$. To our knowledge, our performance guarantees constitute the very first results on this problem setting.'
volume: 125
URL: http://proceedings.mlr.press/v125/neu20b.html
PDF: http://proceedings.mlr.press/v125/neu20b/neu20b.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-neu20b.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Neu
given: Gergely
- family: Olkhovskaya
given: Julia
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 3049-3068
id: neu20b
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 3049
lastpage: 3068
published: 2020-07-15 00:00:00 +0000
- title: 'An $\widetilde\mathcal{O}(m/\varepsilon^3.5)$-Cost Algorithm for Semidefinite Programs with Diagonal Constraints'
abstract: 'We provide a first-order algorithm for semidefinite programs (SDPs) with diagonal constraints on the matrix variable. Our algorithm outputs an $\varepsilon$-optimal solution with a run time of $\widetilde{\mathcal{O}}(m/\varepsilon^{3.5})$, where $m$ is the number of non-zero entries in the cost matrix. This improves upon the previous best run time of $\widetilde{\mathcal{O}}(m/\varepsilon^{4.5})$ by Arora and Kale. As a corollary of our result, given an instance of the Max-Cut problem with $n$ vertices and $m \gg n$ edges, our algorithm returns a $(1 - \varepsilon)\alpha_{GW}$ cut in the faster time of $\widetilde{\mathcal{O}}(m/\varepsilon^{3.5})$, where $\alpha_{GW} \approx 0.878567$ is the approximation ratio by Goemans and Williamson. Our key technical contribution is to combine an approximate variant of the Arora-Kale framework of mirror descent for SDPs with the idea of trading off exact computations in every iteration for variance-reduced estimations in most iterations, only periodically resetting the accumulated error with exact computations. This idea, along with the constructed estimator, are of possible independent interest for other problems that use the mirror descent framework. '
volume: 125
URL: http://proceedings.mlr.press/v125/lee20c.html
PDF: http://proceedings.mlr.press/v125/lee20c/lee20c.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-lee20c.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Lee
given: Yin Tat
- family: Padmanabhan
given: Swati
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 3069-3119
id: lee20c
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 3069
lastpage: 3119
published: 2020-07-15 00:00:00 +0000
- title: 'Costly Zero Order Oracles'
abstract: ' We study optimization with an approximate zero order oracle where there is a cost $c(\epsilon)$ associated with querying the oracle with $\epsilon$ accuracy. We consider the task of reconstructing a linear function: given a linear function $f : X \subseteq \R^d \rightarrow [-1,1]$ defined on a not-necessarily-convex set $X$, the goal is to reconstruct $\hat f$ such that $\vert{f(x) - \hat f(x)}\vert \leq \epsilon$ for all $x \in X$. We show that this can be done with cost $O(d \cdot c(\epsilon/d))$. The algorithm is based on a (poly-time computable) John-like theorem for simplices instead of ellipsoids, which may be of independent interest. This question is motivated by optimization with threshold queries, which are common in economic applications such as pricing. With threshold queries, approximating a number up to precision $\epsilon$ requires $c(\epsilon) = \log(1/\epsilon)$. For this, our algorithm has cost $O(d \log(d/\epsilon))$ which matches the $\Omega(d \log(1/\epsilon))$ lower bound up to log factors. '
volume: 125
URL: http://proceedings.mlr.press/v125/paes-leme20a.html
PDF: http://proceedings.mlr.press/v125/paes-leme20a/paes-leme20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-paes-leme20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Paes Leme
given: Renato
- family: Schneider
given: Jon
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 3120-3132
id: paes-leme20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 3120
lastpage: 3132
published: 2020-07-15 00:00:00 +0000
- title: 'Adaptive Submodular Maximization under Stochastic Item Costs'
abstract: ' Constrained maximization of non-decreasing utility functions with submodularity-like properties is at the core of several AI and ML applications including viral marketing, pool-based active learning, adaptive feature selection, and sensor deployment. In this work, we develop adaptive policies for maximizing such functions when both the utility function and item costs may be stochastic. First, we study maximization of an adaptive weak submodular function which is submodular in an approximate and probabilistic sense, subject to a stochastic fractional knapsack constraint, which requires total expected item cost to be within a given capacity. We present the $\beta$-GREEDY policy for this problem; our approximation guarantee for it recovers many known greedy maximization guarantees as special cases. Next, we study maximization of an adaptive submodular function, which is submodular in a probabilistic sense, subject to a stochastic knapsack constraint, which requires the total item cost to be within a given capacity with probability one. We present the MIX policy for this problem; our approximation guarantee for it is the first known approximation guarantee for maximizing a non-linear utility function subject to a stochastic knapsack constraint. Using alternative parameterizations of MIX, we also derive the first known approximation guarantees for maximizing an adaptive submodular function subject to a deterministic knapsack constraint. Our guarantees are powered by an innovative differential analysis technique which models the $\beta$-GREEDY policy using a continuous-time stochastic reward process of a particle whose reward is related to the optimal utility through a differential inequality. The solution to this inequality yields our $\beta$-GREEDY guarantee. We combine differential analysis with a variety of other ideas to derive our MIX guarantees.'
volume: 125
URL: http://proceedings.mlr.press/v125/parthasarathy20a.html
PDF: http://proceedings.mlr.press/v125/parthasarathy20a/parthasarathy20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-parthasarathy20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Parthasarathy
given: Srinivasan
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 3133-3151
id: parthasarathy20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 3133
lastpage: 3151
published: 2020-07-15 00:00:00 +0000
- title: 'Covariance-adapting algorithm for semi-bandits with application to sparse outcomes'
abstract: ' We investigate \emph{stochastic combinatorial semi-bandits}, where the entire joint distribution of outcomes impacts the complexity of the problem instance (unlike in the standard bandits). Typical distributions considered depend on specific parameter values, whose prior knowledge is required in theory but quite difficult to estimate in practice; an example is the commonly assumed \emph{sub-Gaussian} family. We alleviate this issue by instead considering a new general family of \emph{sub-exponential} distributions, which contains bounded and Gaussian ones. We prove a new lower bound on the regret on this family, that is parameterized by the \emph{unknown} covariance matrix, a tighter quantity than the sub-Gaussian matrix. We then construct an algorithm that uses covariance estimates, and provide a tight asymptotic analysis of the regret. Finally, we apply and extend our results to the family of sparse outcomes, which has applications in many recommender systems. '
volume: 125
URL: http://proceedings.mlr.press/v125/perrault20a.html
PDF: http://proceedings.mlr.press/v125/perrault20a/perrault20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-perrault20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Perrault
given: Pierre
- family: Valko
given: Michal
- family: Perchet
given: Vianney
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 3152-3184
id: perrault20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 3152
lastpage: 3184
published: 2020-07-15 00:00:00 +0000
- title: 'Finite-Time Analysis of Asynchronous Stochastic Approximation and $Q$-Learning'
abstract: ' We consider a general asynchronous Stochastic Approximation (SA) scheme featuring a weighted infinity-norm contractive operator, and prove a bound on its finite-time convergence rate on a single trajectory. Additionally, we specialize the result to asynchronous $Q$-learning. The resulting bound matches the sharpest available bound for synchronous $Q$-learning, and improves over previous known bounds for asynchronous $Q$-learning.'
volume: 125
URL: http://proceedings.mlr.press/v125/qu20a.html
PDF: http://proceedings.mlr.press/v125/qu20a/qu20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-qu20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Qu
given: Guannan
- family: Wierman
given: Adam
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 3185-3205
id: qu20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 3185
lastpage: 3205
published: 2020-07-15 00:00:00 +0000
- title: 'List Decodable Subspace Recovery'
abstract: ' Learning from data in the presence of outliers is a fundamental problem in statistics. In this work, we study robust statistics in the presence of overwhelming outliers for the fundamental problem of subspace recovery. Given a dataset where an $\alpha$ fraction (less than half) of the data is distributed uniformly in an unknown $k$ dimensional subspace in $d$ dimensions, and with no additional assumptions on the remaining data, the goal is to recover a succinct list of $O(\frac{1}{\alpha})$ subspaces one of which is nontrivially correlated with the planted subspace. We provide the first polynomial time algorithm for the ’list decodable subspace recovery’ problem, and subsume it under a more general framework of list decoding over distributions that are "certifiably resilient" capturing state of the art results for list decodable mean estimation and regression.'
volume: 125
URL: http://proceedings.mlr.press/v125/raghavendra20a.html
PDF: http://proceedings.mlr.press/v125/raghavendra20a/raghavendra20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-raghavendra20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Raghavendra
given: Prasad
- family: Yau
given: Morris
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 3206-3226
id: raghavendra20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 3206
lastpage: 3226
published: 2020-07-15 00:00:00 +0000
- title: 'Tsallis-INF for Decoupled Exploration and Exploitation in Multi-armed Bandits'
abstract: ' We consider a variation of the multi-armed bandit problem, introduced by Avner et al. (2012), in which the forecaster is allowed to choose one arm to explore and one arm to exploit at every round. The loss of the exploited arm is blindly suffered by the forecaster, while the loss of the explored arm is observed without being suffered. The goal of the learner is to minimize the regret. We derive a new algorithm using regularization by Tsallis entropy to achieve best of both worlds guarantees. In the adversarial setting we show that the algorithm achieves the minimax optimal $O(\sqrt{KT})$ regret bound, slightly improving on the result of Avner et al.. In the stochastic regime the algorithm achieves a time-independent regret bound, significantly improving on the result of Avner et al.. The algorithm also achieves the same time-independent regret bound in the more general stochastically constrained adversarial regime introduced by Wei and Luo (2018).'
volume: 125
URL: http://proceedings.mlr.press/v125/rouyer20a.html
PDF: http://proceedings.mlr.press/v125/rouyer20a/rouyer20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-rouyer20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Rouyer
given: Chloé
- family: Seldin
given: Yevgeny
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 3227-3249
id: rouyer20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 3227
lastpage: 3249
published: 2020-07-15 00:00:00 +0000
- title: 'How Good is SGD with Random Shuffling?'
abstract: ' We study the performance of stochastic gradient descent (SGD) on smooth and strongly-convex finite-sum optimization problems. In contrast to the majority of existing theoretical works, which assume that individual functions are sampled with replacement, we focus here on popular but poorly-understood heuristics, which involve going over random permutations of the individual functions. This setting has been investigated in several recent works, but the optimal error rates remain unclear. In this paper, we provide lower bounds on the expected optimization error with these heuristics (using SGD with any constant step size), which elucidate their advantages and disadvantages. In particular, we prove that after $k$ passes over $n$ individual functions, if the functions are re-shuffled after every pass, the best possible optimization error for SGD is at least $\Omega\left(1/(nk)^2+1/nk^3\right)$, which partially corresponds to recently derived upper bounds. Moreover, if the functions are only shuffled once, then the lower bound increases to $\Omega(1/nk^2)$. Since there are strictly smaller upper bounds for repeated reshuffling, this proves an inherent performance gap between SGD with single shuffling and repeated shuffling. As a more minor contribution, we also provide a non-asymptotic $\Omega(1/k^2)$ lower bound (independent of $n$) for the incremental gradient method, when no random shuffling takes place. Finally, we provide an indication that our lower bounds are tight, by proving matching upper bounds for univariate quadratic functions.'
volume: 125
URL: http://proceedings.mlr.press/v125/safran20a.html
PDF: http://proceedings.mlr.press/v125/safran20a/safran20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-safran20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Safran
given: Itay
- family: Shamir
given: Ohad
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 3250-3284
id: safran20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 3250
lastpage: 3284
published: 2020-07-15 00:00:00 +0000
- title: 'A Nearly Optimal Variant of the Perceptron Algorithm for the Uniform Distribution on the Unit Sphere'
abstract: ' We show a simple perceptron-like algorithm to learn origin-centered halfspaces in $\mathbb{R}^n$ with accuracy $1-\epsilon$ and confidence $1-\delta$ in time $\mathcal{O}\left(\frac{n^2}{\epsilon}\left(\log \frac{1}{\epsilon}+\log \frac{1}{\delta}\right)\right)$ using $\mathcal{O}\left(\frac{n}{\epsilon}\left(\log \frac{1}{\epsilon}+\log \frac{1}{\delta}\right)\right)$ labeled examples drawn uniformly from the unit $n$-sphere. This improves upon algorithms given in Baum(1990), Long(1994) and Servedio(1999). The time and sample complexity of our algorithm match the lower bounds given in Long(1995) up to logarithmic factors. '
volume: 125
URL: http://proceedings.mlr.press/v125/schmalhofer20a.html
PDF: http://proceedings.mlr.press/v125/schmalhofer20a/schmalhofer20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-schmalhofer20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Schmalhofer
given: Marco
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 3285-3295
id: schmalhofer20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 3285
lastpage: 3295
published: 2020-07-15 00:00:00 +0000
- title: 'Logistic Regression Regret: What’s the Catch?'
abstract: ' We address the problem of the achievable regret rates with online logistic regression. We derive lower bounds with logarithmic regret under $L_1$, $L_2$, and $L_\infty$ constraints on the parameter values. The bounds are dominated by $d/2 \log T$, where $T$ is the horizon and $d$ is the dimensionality of the parameter space. We show their achievability for $d=o(T^{1/3})$ in all these cases with Bayesian methods, that achieve them up to a $d/2 \log d$ term. Interesting different behaviors are shown for larger dimensionality. Specifically, on the negative side, if $d = \Omega(\sqrt{T})$, any algorithm is guaranteed regret of $\Omega(d \log T)$ (greater than $\Theta(\sqrt{T})$) under $L_\infty$ constraints on the parameters (and the example features). On the positive side, under $L_1$ constraints on the parameters, there exist Bayesian algorithms that can achieve regret that is sub-linear in $d$ for the asymptotically larger values of $d$. For $L_2$ constraints, it is shown that for large enough $d$, the regret remains linear in $d$ but no longer logarithmic in $T$. Adapting the \emph{redundancy-capacity\/} theorem from information theory, we demonstrate a principled methodology based on grids of parameters to derive lower bounds. Grids are also utilized to derive some upper bounds. Our results strengthen results by Kakade and Ng (2005) and Foster et al. (2018) for upper bounds for this problem, introduce novel lower bounds, and adapt a methodology that can be used to obtain such bounds for other related problems. They also give a novel characterization of the asymptotic behavior when the dimension of the parameter space is allowed to grow with $T$. They additionally strengthen connections to the information theory literature, demonstrating that the actual regret for logistic regression depends on the richness of the parameter class, where even within this problem, richer classes lead to greater regret.'
volume: 125
URL: http://proceedings.mlr.press/v125/shamir20a.html
PDF: http://proceedings.mlr.press/v125/shamir20a/shamir20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-shamir20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Shamir
given: Gil I
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 3296-3319
id: shamir20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 3296
lastpage: 3319
published: 2020-07-15 00:00:00 +0000
- title: 'Improper Learning for Non-Stochastic Control'
abstract: ' We consider the problem of controlling a possibly unknown linear dynamical system with adversarial perturbations, adversarially chosen convex loss functions, and partially observed states, known as non-stochastic control. We introduce a controller parametrization based on the denoised observations, and prove that applying online gradient descent to this parametrization yields a new controller which attains sublinear regret vs. a large class of closed-loop policies. In the fully-adversarial setting, our controller attains an optimal regret bound of $\sqrt{T}$-when the system is known, and, when combined with an initial stage of least-squares estimation, $T^{2/3}$ when the system is unknown; both yield the first sublinear regret for the partially observed setting. Our bounds are the first in the non-stochastic control setting that compete with \emph{all} stabilizing linear dynamical controllers, not just state feedback. Moreover, in the presence of semi-adversarial noise containing both stochastic and adversarial components, our controller attains the optimal regret bounds of $\mathrm{poly}(\log T)$ when the system is known, and $\sqrt{T}$ when unknown. To our knowledge, this gives the first end-to-end $\sqrt{T}$ regret for online Linear Quadratic Gaussian controller, and applies in a more general setting with adversarial losses and semi-adversarial noise.'
volume: 125
URL: http://proceedings.mlr.press/v125/simchowitz20a.html
PDF: http://proceedings.mlr.press/v125/simchowitz20a/simchowitz20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-simchowitz20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Simchowitz
given: Max
- family: Singh
given: Karan
- family: Hazan
given: Elad
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 3320-3436
id: simchowitz20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 3320
lastpage: 3436
published: 2020-07-15 00:00:00 +0000
- title: 'Reasoning About Generalization via Conditional Mutual Information'
abstract: ' We provide an information-theoretic framework for studying the generalization properties of machine learning algorithms. Our framework ties together existing approaches, including uniform convergence bounds and recent methods for adaptive data analysis. Specifically, we use Conditional Mutual Information (CMI) to quantify how well the input (i.e., the training data) can be recognized given the output (i.e., the trained model) of the learning algorithm. We show that bounds on CMI can be obtained from VC dimension, compression schemes, differential privacy, and other methods. We then show that bounded CMI implies various forms of generalization.'
volume: 125
URL: http://proceedings.mlr.press/v125/steinke20a.html
PDF: http://proceedings.mlr.press/v125/steinke20a/steinke20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-steinke20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Steinke
given: Thomas
- family: Zakynthinou
given: Lydia
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 3437-3452
id: steinke20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 3437
lastpage: 3452
published: 2020-07-15 00:00:00 +0000
- title: 'Estimation and Inference with Trees and Forests in High Dimensions'
abstract: ' Regression Trees [Breiman et al. 1984] and Random Forests [Breiman 2001], are one of the most widely used estimation methods by machine learning practitioners. Despite their widespread use, their theoretical underpinnings are far from being fully understood. Recent breakthrough advances have shown that such greedily built trees are asymptotically consistent [Biau et al. 2010, Denil et al. 2014, Scornet et al. 2015] in the low dimensional regime, where the number of features is a constant, independent of the sample size. Also, the works of [Mentch et al. 2016, Wager et al. 2018] provide asymptotic normality results for honest versions of Random Forests. In this work, we analyze the performance of regression trees and forests with binary features in the high-dimensional regime, where the number of features can grow exponentially with the number of samples. We show that trees and forests built greedily based on the celebrated CART criterion, provably adapt to sparsity: when only a subset $R$, of size $r$, of the features are relevant, then the mean squared error of appropriately shallow trees, or fully grown honest forests, scales exponentially only with the number of relevant features and depends only logarithmically on the overall number of features. More precisely, we identify three regimes, each providing different dependence on the number of relevant features. When the relevant variables are “weakly” relevant (in the sense that there is not strong separation between the relevant and irrelevant variables in terms of their ability to reduce variance), then shallow trees achieve “slow rates” on the mean squared error of the order of $2^r/\sqrt{n}$, when variables are independent, and $1/n^{1/(r+2)}$, when variables are dependent. When the relevant variables are “strongly” relevant, in that there is a separation in their ability to reduce variance as compared to the irrelevant ones, by a constant $\beta_{\min}$, then we show that greedily built shallow trees and fully grown honest forests can achieve fast parametric mean squared error rates of the order of $2^r/(\beta_{\min}\,{n})$. When variables are strongly relevant, we also show that the predictions of sub-sampled honest forests have an asymptotically normal distribution centered around their true values and whose variance scales at most as $O(2^r \log(n)/(\beta_{\min}\,{n}))$. Thus, sub-sampled honest forests are provably a data-adaptive method for non-parametric inference, that adapts to the latent sparsity dimension of the data generating distribution, as opposed to classical non-parametric regression approaches. Our results show that, at least for the case of binary features, forest based algorithms can offer immense improvement on the statistical power of non-parametric hypothesis tests in high-dimensional regimes. '
volume: 125
URL: http://proceedings.mlr.press/v125/syrgkanis20a.html
PDF: http://proceedings.mlr.press/v125/syrgkanis20a/syrgkanis20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-syrgkanis20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Syrgkanis
given: Vasilis
- family: Zampetakis
given: Manolis
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 3453-3454
id: syrgkanis20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 3453
lastpage: 3454
published: 2020-07-15 00:00:00 +0000
- title: 'Balancing Gaussian vectors in high dimension'
abstract: ' Motivated by problems in controlled experiments, we study the discrepancy of random matrices with continuous entries where the number of columns $n$ is much larger than the number of rows $m$. Our first result shows that if $\omega(1) = m = o(n)$, a matrix with i.i.d. standard Gaussian entries has discrepancy $\Theta(\sqrt{n} \, 2^{-n/m})$ with high probability. This provides sharp guarantees for Gaussian discrepancy in a regime that had not been considered before in the existing literature. Our results also apply to a more general family of random matrices with continuous i.i.d. entries, assuming that $m = O(n/\log{n})$. The proof is non-constructive and is an application of the second moment method. Our second result is algorithmic and applies to random matrices whose entries are i.i.d. and have a Lipschitz density. We present a randomized polynomial-time algorithm that achieves discrepancy $e^{-\Omega(\log^2(n)/m)}$ with high probability, provided that $m = O(\sqrt{\log{n}})$. In the one-dimensional case, this matches the best known algorithmic guarantees due to Karmarkar–Karp. For higher dimensions $2 \leq m = O(\sqrt{\log{n}})$, this establishes the first efficient algorithm achieving discrepancy smaller than $O( \sqrt{m} )$. '
volume: 125
URL: http://proceedings.mlr.press/v125/turner20a.html
PDF: http://proceedings.mlr.press/v125/turner20a/turner20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-turner20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Turner
given: Paxton
- family: Meka
given: Raghu
- family: Rigollet
given: Philippe
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 3455-3486
id: turner20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 3455
lastpage: 3486
published: 2020-07-15 00:00:00 +0000
- title: 'Active Learning for Identification of Linear Dynamical Systems'
abstract: ' We propose an algorithm to actively estimate the parameters of a linear dynamical system. Given complete control over the system’s input, our algorithm adaptively chooses the inputs to accelerate estimation. We show a finite time bound quantifying the estimation rate our algorithm attains and prove matching upper and lower bounds which guarantee its asymptotic optimality, up to constants. In addition, we show that this optimal rate is unattainable when using Gaussian noise to excite the system, even with optimally tuned covariance, and analyze several examples where our algorithm provably improves over rates obtained by playing noise. Our analysis critically relies on a novel result quantifying the error in estimating the parameters of a dynamical system when arbitrary periodic inputs are being played. We conclude with numerical examples that illustrate the effectiveness of our algorithm in practice. '
volume: 125
URL: http://proceedings.mlr.press/v125/wagenmaker20a.html
PDF: http://proceedings.mlr.press/v125/wagenmaker20a/wagenmaker20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-wagenmaker20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Wagenmaker
given: Andrew
- family: Jamieson
given: Kevin
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 3487-3582
id: wagenmaker20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 3487
lastpage: 3582
published: 2020-07-15 00:00:00 +0000
- title: 'Taking a hint: How to leverage loss predictors in contextual bandits?'
abstract: ' We initiate the study of learning in contextual bandits with the help of loss predictors. The main question we address is whether one can improve over the minimax regret $\mathcal{O}(\sqrt{T})$ for learning over $T$ rounds, when the total error of the predicted losses relative to the realized losses, denoted as $\mathcal{E} \leq T$, is relatively small. We provide a complete answer to this question, with upper and lower bounds for various settings: adversarial and stochastic environments, known and unknown $\mathcal{E}$, and single and multiple predictors. We show several surprising results, such as 1) the optimal regret is $\mathcal{O}(\min\{\sqrt{T}, \sqrt{\mathcal{E}}T^\frac{1}{4}\})$ when $\mathcal{E}$ is known, in contrast to the standard and better bound $\mathcal{O}(\sqrt{\mathcal{E}})$ for non-contextual problems (such as multi-armed bandits); 2) the same bound cannot be achieved if $\mathcal{E}$ is unknown, but as a remedy, $\mathcal{O}(\sqrt{\mathcal{E}}T^\frac{1}{3})$ is achievable; 3) with $M$ predictors, a linear dependence on $M$ is necessary, even though logarithmic dependence is possible for non-contextual problems. We also develop several novel algorithmic techniques to achieve matching upper bounds, including 1) a key \emph{action remapping} technique for optimal regret with known $\mathcal{E}$, 2) computationally efficient implementation of Catoni’s robust mean estimator via an ERM oracle in the stochastic setting with optimal regret, 3) an underestimator for $\mathcal{E}$ via estimating the histogram with bins of exponentially increasing size for the stochastic setting with unknown $\mathcal{E}$, and 4) a self-referential scheme for learning with multiple predictors, all of which might be of independent interest. '
volume: 125
URL: http://proceedings.mlr.press/v125/wei20a.html
PDF: http://proceedings.mlr.press/v125/wei20a/wei20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-wei20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Wei
given: Chen-Yu
- family: Luo
given: Haipeng
- family: Agarwal
given: Alekh
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 3583-3634
id: wei20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 3583
lastpage: 3634
published: 2020-07-15 00:00:00 +0000
- title: 'Kernel and Rich Regimes in Overparametrized Models'
abstract: ' A recent line of work studies overparametrized neural networks in the “kernel regime,” i.e. when during training the network behaves as a kernelized linear predictor, and thus, training with gradient descent has the effect of finding the corresponding minimum RKHS norm solution. This stands in contrast to other studies which demonstrate how gradient descent on overparametrized networks can induce rich implicit biases that are not RKHS norms. Building on an observation by \citet{chizat2018note}, we show how the \textbf{\textit{scale of the initialization}} controls the transition between the “kernel” (aka lazy) and “rich” (aka active) regimes and affects generalization properties in multilayer homogeneous models. We provide a complete and detailed analysis for a family of simple depth-$D$ linear networks that exhibit an interesting and meaningful transition between the kernel and rich regimes, and highlight an interesting role for the \emph{width} of the models. We further demonstrate this transition empirically for matrix factorization and multilayer non-linear networks.'
volume: 125
URL: http://proceedings.mlr.press/v125/woodworth20a.html
PDF: http://proceedings.mlr.press/v125/woodworth20a/woodworth20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-woodworth20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Woodworth
given: Blake
- family: Gunasekar
given: Suriya
- family: Lee
given: Jason D.
- family: Moroshko
given: Edward
- family: Savarese
given: Pedro
- family: Golan
given: Itay
- family: Soudry
given: Daniel
- family: Srebro
given: Nathan
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 3635-3673
id: woodworth20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 3635
lastpage: 3673
published: 2020-07-15 00:00:00 +0000
- title: 'Learning Zero-Sum Simultaneous-Move Markov Games Using Function Approximation and Correlated Equilibrium'
abstract: ' In this work, we develop provably efficient reinforcement learning algorithms for two-player zero-sum Markov games with simultaneous moves. We consider a family of Markov games where the reward function and transition kernel possess a linear structure. Two settings are studied: In the offline setting, we control both players and the goal is to find the Nash Equilibrium efficiently by minimizing the worst-case duality gap. In the online setting, we control a single player and play against an arbitrary opponent; the goal is to minimize the regret. For both settings, we propose an optimistic variant of the least-squares minimax value iteration algorithm. We show that our algorithm is computationally efficient and provably achieves an $\tilde O(\sqrt{d^3 H^3 T} )$ upper bound on the duality gap and regret, without requiring additional assumptions on the sampling model. We highlight that our setting requires overcoming several new challenges that are absent in MDPs or turn-based Markov games. In particular, to achieve optimism under the simultaneous-move games, we construct both upper and lower confidence bounds of the value function, and then derive the optimistic policy by solving a general-sum matrix game with these bounds as the payoff matrices. As finding the Nash Equilibrium of this general-sum game is computationally hard, our algorithm instead solves for a Coarse Correlated Equilibrium (CCE), which can be obtained efficiently via linear programming. To our best knowledge, such a CCE-based mechanism for implementing optimism has not appeared in the literature and might be of interest in its own right.'
volume: 125
URL: http://proceedings.mlr.press/v125/xie20a.html
PDF: http://proceedings.mlr.press/v125/xie20a/xie20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-xie20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Xie
given: Qiaomin
- family: Chen
given: Yudong
- family: Wang
given: Zhaoran
- family: Yang
given: Zhuoran
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 3674-3682
id: xie20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 3674
lastpage: 3682
published: 2020-07-15 00:00:00 +0000
- title: 'Tree-projected gradient descent for estimating gradient-sparse parameters on graphs'
abstract: ' We study estimation of a gradient-sparse parameter vector $\boldsymbol{\theta}^* \in \mathbb{R}^p$, having strong gradient-sparsity $s^*:=\|\nabla_G \boldsymbol{\theta}^*\|_0$ on an underlying graph $G$. Given observations $Z_1,\ldots,Z_n$ and a smooth, convex loss function $\mathcal{L}$ for which $\boldsymbol{\theta}^*$ minimizes the population risk $\mathbb{E}[\mathcal{L}(\boldsymbol{\theta};Z_1,\ldots,Z_n)]$, we propose to estimate $\boldsymbol{\theta}^*$ by a projected gradient descent algorithm that iteratively and approximately projects gradient steps onto spaces of vectors having small gradient-sparsity over low-degree spanning trees of $G$. We show that, under suitable restricted strong convexity and smoothness assumptions for the loss, the resulting estimator achieves the squared-error risk $\frac{s^*}{n} \log (1+\frac{p}{s^*})$ up to a multiplicative constant that is independent of $G$. In contrast, previous polynomial-time algorithms have only been shown to achieve this guarantee in more specialized settings, or under additional assumptions for $G$ and/or the sparsity pattern of $\nabla_G \boldsymbol{\theta}^*$. As applications of our general framework, we apply our results to the examples of linear models and generalized linear models with random design.'
volume: 125
URL: http://proceedings.mlr.press/v125/xu20a.html
PDF: http://proceedings.mlr.press/v125/xu20a/xu20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-xu20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Xu
given: Sheng
- family: Fan
given: Zhou
- family: Negahban
given: Sahand
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 3683-3708
id: xu20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 3683
lastpage: 3708
published: 2020-07-15 00:00:00 +0000
- title: 'Non-asymptotic Analysis for Nonparametric Testing'
abstract: ' We develop a non-asymptotic framework for hypothesis testing in nonparametric regression where the true regression function belongs to a Sobolev space. Our statistical guarantees are exact in the sense that Type I and II errors are controlled for any finite sample size. Meanwhile, one proposed test is shown to achieve minimax rate optimality in the asymptotic sense. An important consequence of this non-asymptotic theory is a new and practically useful formula for selecting the optimal smoothing parameter in the testing statistic. Extensions of our results to general reproducing kernel Hilbert spaces and non-Gaussian error regression are also discussed.'
volume: 125
URL: http://proceedings.mlr.press/v125/yang20a.html
PDF: http://proceedings.mlr.press/v125/yang20a/yang20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-yang20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Yang
given: Yun
- family: Shang
given: Zuofeng
- family: Cheng
given: Guang
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 3709-3755
id: yang20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 3709
lastpage: 3755
published: 2020-07-15 00:00:00 +0000
- title: 'Learning a Single Neuron with Gradient Methods'
abstract: ' We consider the fundamental problem of learning a single neuron $\mathbf{x}\mapsto \sigma(\mathbf{w}^\top\mathbf{x})$ in a realizable setting, using standard gradient methods with random initialization, and under general families of input distributions and activations. On the one hand, we show that some assumptions on both the distribution and the activation function are necessary. On the other hand, we prove positive guarantees under mild assumptions, which go significantly beyond those studied in the literature so far. We also point out and study the challenges in further strengthening and generalizing our results.'
volume: 125
URL: http://proceedings.mlr.press/v125/yehudai20a.html
PDF: http://proceedings.mlr.press/v125/yehudai20a/yehudai20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-yehudai20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Yehudai
given: Gilad
- family: Ohad
given: Shamir
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 3756-3786
id: yehudai20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 3756
lastpage: 3786
published: 2020-07-15 00:00:00 +0000
- title: 'Nearly Non-Expansive Bounds for Mahalanobis Hard Thresholding'
abstract: ' Given a vector $w \in \mathbb{R}^p$ and a positive semi-definite matrix $A \in \mathbb{R}^{p\times p}$, we study the expansion ratio bound for the following defined Mahalanobis hard thresholding operator of $w$: \[ \mathcal{H}_{A,k}(w):=\argmin_{\|\theta\|_0\le k} \frac{1}{2}\|\theta - w\|^2_A, \]{where} $k\le p$ is the desired sparsity level. The core contribution of this paper is to prove that for any $\bar k$-sparse vector $\bar w$ with $\bar k < k$, the estimation error $\|\mathcal{H}_{A,k}(w) - \bar w\|_A$ satisfies \[ \|\mathcal{H}_{A,k}(w) - \bar w\|^2_A \le \left(1+ \mathcal{O}\left(\kappa(A,2k) \sqrt{\frac{\bar k }{k - \bar k}}\right)\right) \|{w} - \bar w\|^2_A, \]{where} $\kappa(A,2k)$ is the restricted strong condition number of $A$ over $(2k)$-sparse subspace. This estimation error bound is nearly non-expansive when $k$ is sufficiently larger than $\bar k$. Specially when $A$ is the identity matrix such that $\kappa(A,2k)\equiv1$, our bound recovers the previously known nearly non-expansive bounds for Euclidean hard thresholding operator. We further show that such a bound extends to an approximate version of $\mathcal{H}_{A,k}(w)$ estimated by Hard Thresholding Pursuit (HTP) algorithm. We demonstrate the applicability of these bounds to the mean squared error analysis of HTP and its novel extension based on preconditioning method. Numerical evidence is provided to support our theory and demonstrate the superiority of the proposed preconditioning HTP algorithm. '
volume: 125
URL: http://proceedings.mlr.press/v125/yuan20a.html
PDF: http://proceedings.mlr.press/v125/yuan20a/yuan20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-yuan20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Yuan
given: Xiao-Tong
- family: Li
given: Ping
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 3787-3813
id: yuan20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 3787
lastpage: 3813
published: 2020-07-15 00:00:00 +0000
- title: 'Wasserstein Control of Mirror Langevin Monte Carlo'
abstract: ' Discretized Langevin diffusions are efficient Monte Carlo methods for sampling from high dimensional target densities that are log-Lipschitz-smooth and (strongly) log-concave. In particular, the Euclidean Langevin Monte Carlo sampling algorithm has received much attention lately, leading to a detailed understanding of its non-asymptotic convergence properties and of the role that smoothness and log-concavity play in the convergence rate. Distributions that do not possess these regularity properties can be addressed by considering a Riemannian Langevin diffusion with a metric capturing the local geometry of the log-density. However, the Monte Carlo algorithms derived from discretizations of such Riemannian Langevin diffusions are notoriously difficult to analyze. In this paper, we consider Langevin diffusions on a Hessian-type manifold and study a discretization that is closely related to the mirror-descent scheme. We establish for the first time a non-asymptotic upper-bound on the sampling error of the resulting Hessian Riemannian Langevin Monte Carlo algorithm. This bound is measured according to a Wasserstein distance induced by a Riemannian metric ground cost capturing the squared Hessian structure and closely related to a self-concordance-like condition. The upper-bound implies, for instance, that the iterates contract toward a Wasserstein ball around the target density whose radius is made explicit. Our theory recovers existing Euclidean results and can cope with a wide variety of Hessian metrics related to highly non-flat geometries. '
volume: 125
URL: http://proceedings.mlr.press/v125/zhang20a.html
PDF: http://proceedings.mlr.press/v125/zhang20a/zhang20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-zhang20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Zhang
given: Kelvin Shuangjian
- family: Peyré
given: Gabriel
- family: Fadili
given: Jalal
- family: Pereyra
given: Marcelo
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 3814-3841
id: zhang20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 3814
lastpage: 3841
published: 2020-07-15 00:00:00 +0000
- title: 'Open Problem: Model Selection for Contextual Bandits'
abstract: 'In statistical learning, algorithms for model selection allow the learner to adapt to the complexity of the best hypothesis class in a sequence. We ask whether similar guarantees are possible for contextual bandit learning.'
volume: 125
URL: http://proceedings.mlr.press/v125/foster20a.html
PDF: http://proceedings.mlr.press/v125/foster20a/foster20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-foster20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Foster
given: Dylan J.
- family: Krishnamurthy
given: Akshay
- family: Luo
given: Haipeng
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 3842-3846
id: foster20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 3842
lastpage: 3846
published: 2020-07-15 00:00:00 +0000
- title: 'Open Problem: Tight Convergence of SGD in Constant Dimension'
abstract: 'Stochastic Gradient Descent (SGD) is one of the most popular optimization methods in machine learning and has been studied extensively since the early 50’s. However, our understanding of this fundamental algorithm is still lacking in certain aspects. We point out to a gap that remains between the known upper and lower bounds for the expected suboptimality of the last SGD point whenever the dimension is a constant independent of the number of SGD iterations $T$, and in particular, that the gap is still unaddressed even in the one dimensional case. For the latter, we provide evidence that the correct rate is $\Theta(1/\sqrt{T})$ and conjecture that the same applies in any (constant) dimension.'
volume: 125
URL: http://proceedings.mlr.press/v125/koren20a.html
PDF: http://proceedings.mlr.press/v125/koren20a/koren20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-koren20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Koren
given: Tomer
- family: Segal
given: Shahar
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 3847-3851
id: koren20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 3847
lastpage: 3851
published: 2020-07-15 00:00:00 +0000
- title: 'Open Problem: Average-Case Hardness of Hypergraphic Planted Clique Detection'
abstract: 'We note the significance of hypergraphic planted clique (HPC) detection in the investigation of computational hardness for a range of tensor problems. We ask if more evidence for the computational hardness of HPC detection can be developed. In particular, we conjecture if it is possible to establish the equivalence of the computational hardness between HPC and PC detection.'
volume: 125
URL: http://proceedings.mlr.press/v125/luo20a.html
PDF: http://proceedings.mlr.press/v125/luo20a/luo20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-luo20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Luo
given: Yuetian
- family: Zhang
given: Anru R
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 3852-3856
id: luo20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 3852
lastpage: 3856
published: 2020-07-15 00:00:00 +0000
- title: 'Open Problem: Information Complexity of VC Learning'
abstract: 'Uniform convergence approaches learning by studying the complexity of hypothesis classes. In particular, hypothesis classes with bounded Vapnik-Chervonenkis dimension exhibit strong uniform convergence, that is, any hypothesis in the class has low generalization error. On the other hand, a long line of work studies the information complexity of a learning algorithm, as it is connected to several desired properties, including generalization. We ask whether all VC classes admit a learner with low information complexity which achieves the generalization bounds guaranteed by uniform convergence. Specifically, since we know that this is not possible if we consider proper and consistent learners and measure information complexity in terms of the mutual information (Bassily et al., 2018), we are interested in learners with low information complexity measured in terms of the recently introduced notion of CMI (Steinke and Zakynthinou, 2020). Can we obtain tight bounds on the information complexity of a learning algorithm for a VC class (via CMI), thus exactly retrieving the known generalization bounds implied for this class by uniform convergence?'
volume: 125
URL: http://proceedings.mlr.press/v125/steinke20b.html
PDF: http://proceedings.mlr.press/v125/steinke20b/steinke20b.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-steinke20b.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Steinke
given: Thomas
- family: Zakynthinou
given: Lydia
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 3857-3863
id: steinke20b
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 3857
lastpage: 3863
published: 2020-07-15 00:00:00 +0000
- title: 'Open Problem: Fast and Optimal Online Portfolio Selection'
abstract: 'Online portfolio selection has received much attention in the COLT community since its introduction by Cover, but all state-of-the-art methods fall short in at least one of the following ways: they are either i) computationally infeasible; or ii) they do not guarantee optimal regret; or iii) they assume the gradients are bounded, which is unnecessary and cannot be guaranteed. We are interested in a natural follow-the-regularized-leader (FTRL) approach based on the log barrier regularizer, which is computationally feasible. The open problem we put before the community is to formally prove whether this approach achieves the optimal regret. Resolving this question will likely lead to new techniques to analyse FTRL algorithms. There are also interesting technical connections to self-concordance, which has previously been used in the context of bandit convex optimization.'
volume: 125
URL: http://proceedings.mlr.press/v125/van-erven20a.html
PDF: http://proceedings.mlr.press/v125/van-erven20a/van-erven20a.pdf
edit: https://github.com/mlresearch//v125/edit/gh-pages/_posts/2020-07-15-van-erven20a.md
series: 'Proceedings of Machine Learning Research'
container-title: 'Proceedings of Thirty Third Conference on Learning Theory'
publisher: 'PMLR'
author:
- family: Van Erven
given: Tim
- family: Van der Hoeven
given: Dirk
- family: Kotłowski
given: Wojciech
- family: Koolen
given: Wouter M.
editor:
- family: Abernethy
given: Jacob
- family: Agarwal
given: Shivani
page: 3864-3869
id: van-erven20a
issued:
date-parts:
- 2020
- 7
- 15
firstpage: 3864
lastpage: 3869
published: 2020-07-15 00:00:00 +0000