Exponentially Improved Dimensionality Reduction for l1: Subspace Embeddings and Independence Testing

Yi Li, David Woodruff, Taisuke Yasuda
Proceedings of Thirty Fourth Conference on Learning Theory, PMLR 134:3111-3195, 2021.

Abstract

Despite many applications, dimensionality reduction in the ℓ1-norm is much less understood than in the Euclidean norm. We give two new oblivious dimensionality reduction techniques for the ℓ1-norm which improve exponentially over prior ones: - We design a distribution over random matrices S∈Rr×n, where r=2poly(d/(εδ)), such that given any matrix A∈Rn×d, with probability at least 1−δ, simultaneously for all x, ‖. Note that S is linear, does not depend on A, and maps \ell_1 into \ell_1. Our distribution provides an exponential improvement on the previous best known map of Wang and Woodruff (SODA, 2019), which required r = 2^{2^{\Omega(d)}}, even for constant \varepsilon and \delta. Our bound is optimal, up to a polynomial factor in the exponent, given a known 2^{\textrm{poly}(d)} lower bound for constant \varepsilon and \delta. - We design a distribution over matrices S \in \mathbb{R}^{k \times n}, where k = 2^{O(q^2)}(\varepsilon^{-1} q \log d)^{O(q)}, such that given any q-mode tensor A \in (\mathbb{R}^{d})^{\otimes q}, one can estimate the entrywise \ell_1-norm \|A\|_1 from S(A). Moreover, S = S^1 \otimes S^2 \otimes \cdots \otimes S^q and so given vectors u_1, \ldots, u_q \in \mathbb{R}^d, one can compute S(u_1 \otimes u_2 \otimes \cdots \otimes u_q) in time 2^{O(q^2)}(\varepsilon^{-1} q \log d)^{O(q)}, which is much faster than the d^q time required to form u_1 \otimes u_2 \otimes \cdots \otimes u_q. Our linear map gives a streaming algorithm for independence testing using space 2^{O(q^2)}(\varepsilon^{-1} q \log d)^{O(q)}, improving the previous doubly exponential (\varepsilon^{-1} \log d)^{q^{O(q)}} space bound of Braverman and Ostrovsky (STOC, 2010). For subspace embeddings, we also study the setting when A is itself drawn from distributions with independent entries, and obtain a polynomial embedding dimension. For independence testing, we also give algorithms for any distance measure with a polylogarithmic-sized sketch and satisfying an approximate triangle inequality.

Cite this Paper

BibTeX
@InProceedings{pmlr-v134-li21c, title = {Exponentially Improved Dimensionality Reduction for l1: Subspace Embeddings and Independence Testing}, author = {Li, Yi and Woodruff, David and Yasuda, Taisuke}, booktitle = {Proceedings of Thirty Fourth Conference on Learning Theory}, pages = {3111--3195}, year = {2021}, editor = {Belkin, Mikhail and Kpotufe, Samory}, volume = {134}, series = {Proceedings of Machine Learning Research}, month = {15--19 Aug}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v134/li21c/li21c.pdf}, url = {https://proceedings.mlr.press/v134/li21c.html}, abstract = {Despite many applications, dimensionality reduction in the $\ell_1$-norm is much less understood than in the Euclidean norm. We give two new oblivious dimensionality reduction techniques for the $\ell_1$-norm which improve exponentially over prior ones: - We design a distribution over random matrices $S \in \mathbb{R}^{r \times n}$, where $r = 2^{\textrm{poly}(d/(\varepsilon \delta))}$, such that given any matrix $A \in \mathbb{R}^{n \times d}$, with probability at least $1-\delta$, simultaneously for all $x$, $\|SAx\|_1 = (1 \pm \varepsilon)\|Ax\|_1$. Note that $S$ is linear, does not depend on $A$, and maps $\ell_1$ into $\ell_1$. Our distribution provides an exponential improvement on the previous best known map of Wang and Woodruff (SODA, 2019), which required $r = 2^{2^{\Omega(d)}}$, even for constant $\varepsilon$ and $\delta$. Our bound is optimal, up to a polynomial factor in the exponent, given a known $2^{\textrm{poly}(d)}$ lower bound for constant $\varepsilon$ and $\delta$. - We design a distribution over matrices $S \in \mathbb{R}^{k \times n}$, where $k = 2^{O(q^2)}(\varepsilon^{-1} q \log d)^{O(q)}$, such that given any $q$-mode tensor $A \in (\mathbb{R}^{d})^{\otimes q}$, one can estimate the entrywise $\ell_1$-norm $\|A\|_1$ from $S(A)$. Moreover, $S = S^1 \otimes S^2 \otimes \cdots \otimes S^q$ and so given vectors $u_1, \ldots, u_q \in \mathbb{R}^d$, one can compute $S(u_1 \otimes u_2 \otimes \cdots \otimes u_q)$ in time $2^{O(q^2)}(\varepsilon^{-1} q \log d)^{O(q)}$, which is much faster than the $d^q$ time required to form $u_1 \otimes u_2 \otimes \cdots \otimes u_q$. Our linear map gives a streaming algorithm for independence testing using space $2^{O(q^2)}(\varepsilon^{-1} q \log d)^{O(q)}$, improving the previous doubly exponential $(\varepsilon^{-1} \log d)^{q^{O(q)}}$ space bound of Braverman and Ostrovsky (STOC, 2010). For subspace embeddings, we also study the setting when $A$ is itself drawn from distributions with independent entries, and obtain a polynomial embedding dimension. For independence testing, we also give algorithms for any distance measure with a polylogarithmic-sized sketch and satisfying an approximate triangle inequality.} }
Endnote
%0 Conference Paper %T Exponentially Improved Dimensionality Reduction for l1: Subspace Embeddings and Independence Testing %A Yi Li %A David Woodruff %A Taisuke Yasuda %B Proceedings of Thirty Fourth Conference on Learning Theory %C Proceedings of Machine Learning Research %D 2021 %E Mikhail Belkin %E Samory Kpotufe %F pmlr-v134-li21c %I PMLR %P 3111--3195 %U https://proceedings.mlr.press/v134/li21c.html %V 134 %X Despite many applications, dimensionality reduction in the $\ell_1$-norm is much less understood than in the Euclidean norm. We give two new oblivious dimensionality reduction techniques for the $\ell_1$-norm which improve exponentially over prior ones: - We design a distribution over random matrices $S \in \mathbb{R}^{r \times n}$, where $r = 2^{\textrm{poly}(d/(\varepsilon \delta))}$, such that given any matrix $A \in \mathbb{R}^{n \times d}$, with probability at least $1-\delta$, simultaneously for all $x$, $\|SAx\|_1 = (1 \pm \varepsilon)\|Ax\|_1$. Note that $S$ is linear, does not depend on $A$, and maps $\ell_1$ into $\ell_1$. Our distribution provides an exponential improvement on the previous best known map of Wang and Woodruff (SODA, 2019), which required $r = 2^{2^{\Omega(d)}}$, even for constant $\varepsilon$ and $\delta$. Our bound is optimal, up to a polynomial factor in the exponent, given a known $2^{\textrm{poly}(d)}$ lower bound for constant $\varepsilon$ and $\delta$. - We design a distribution over matrices $S \in \mathbb{R}^{k \times n}$, where $k = 2^{O(q^2)}(\varepsilon^{-1} q \log d)^{O(q)}$, such that given any $q$-mode tensor $A \in (\mathbb{R}^{d})^{\otimes q}$, one can estimate the entrywise $\ell_1$-norm $\|A\|_1$ from $S(A)$. Moreover, $S = S^1 \otimes S^2 \otimes \cdots \otimes S^q$ and so given vectors $u_1, \ldots, u_q \in \mathbb{R}^d$, one can compute $S(u_1 \otimes u_2 \otimes \cdots \otimes u_q)$ in time $2^{O(q^2)}(\varepsilon^{-1} q \log d)^{O(q)}$, which is much faster than the $d^q$ time required to form $u_1 \otimes u_2 \otimes \cdots \otimes u_q$. Our linear map gives a streaming algorithm for independence testing using space $2^{O(q^2)}(\varepsilon^{-1} q \log d)^{O(q)}$, improving the previous doubly exponential $(\varepsilon^{-1} \log d)^{q^{O(q)}}$ space bound of Braverman and Ostrovsky (STOC, 2010). For subspace embeddings, we also study the setting when $A$ is itself drawn from distributions with independent entries, and obtain a polynomial embedding dimension. For independence testing, we also give algorithms for any distance measure with a polylogarithmic-sized sketch and satisfying an approximate triangle inequality.
APA
Li, Y., Woodruff, D. & Yasuda, T.. (2021). Exponentially Improved Dimensionality Reduction for l1: Subspace Embeddings and Independence Testing. Proceedings of Thirty Fourth Conference on Learning Theory, in Proceedings of Machine Learning Research 134:3111-3195 Available from https://proceedings.mlr.press/v134/li21c.html.

Related Material