Comparison Based Learning from Weak Oracles

Ehsan Kazemi, Lin Chen, Sanjoy Dasgupta, Amin Karbasi
Proceedings of the Twenty-First International Conference on Artificial Intelligence and Statistics, PMLR 84:1849-1858, 2018.

Abstract

There is increasing interest in learning algorithms that involve interaction between hu- man and machine. Comparison-based queries are among the most natural ways to get feed- back from humans. A challenge in designing comparison-based interactive learning algorithms is coping with noisy answers. The most common fix is to submit a query several times, but this is not applicable in many situations due to its prohibitive cost and due to the unrealistic assumption of independent noise in different repetitions of the same query. In this paper, we introduce a new weak oracle model, where a non-malicious user responds to a pairwise comparison query only when she is quite sure about the answer. This model is able to mimic the behavior of a human in noise-prone regions. We also consider the ap- plication of this weak oracle model to the problem of content search (a variant of the nearest neighbor search problem) through comparisons. More specifically, we aim at devising efficient algorithms to locate a target object in a database equipped with a dissimilarity metric via invocation of the weak comparison oracle. We propose two algorithms termed Worcs-I and Worcs-II (Weak-Oracle Comparison- based Search), which provably locate the tar- get object in a number of comparisons close to the entropy of the target distribution. While Worcs-I provides better theoretical guarantees, Worcs-II is applicable to more technically challenging scenarios where the algorithm has limited access to the ranking dis- similarity between objects. A series of experiments validate the performance of our proposed algorithms.

Cite this Paper

BibTeX
@InProceedings{pmlr-v84-kazemi18a, title = {Comparison Based Learning from Weak Oracles}, author = {Kazemi, Ehsan and Chen, Lin and Dasgupta, Sanjoy and Karbasi, Amin}, booktitle = {Proceedings of the Twenty-First International Conference on Artificial Intelligence and Statistics}, pages = {1849--1858}, year = {2018}, editor = {Storkey, Amos and Perez-Cruz, Fernando}, volume = {84}, series = {Proceedings of Machine Learning Research}, month = {09--11 Apr}, publisher = {PMLR}, pdf = {http://proceedings.mlr.press/v84/kazemi18a/kazemi18a.pdf}, url = {https://proceedings.mlr.press/v84/kazemi18a.html}, abstract = {There is increasing interest in learning algorithms that involve interaction between hu- man and machine. Comparison-based queries are among the most natural ways to get feed- back from humans. A challenge in designing comparison-based interactive learning algorithms is coping with noisy answers. The most common fix is to submit a query several times, but this is not applicable in many situations due to its prohibitive cost and due to the unrealistic assumption of independent noise in different repetitions of the same query. In this paper, we introduce a new weak oracle model, where a non-malicious user responds to a pairwise comparison query only when she is quite sure about the answer. This model is able to mimic the behavior of a human in noise-prone regions. We also consider the ap- plication of this weak oracle model to the problem of content search (a variant of the nearest neighbor search problem) through comparisons. More specifically, we aim at devising efficient algorithms to locate a target object in a database equipped with a dissimilarity metric via invocation of the weak comparison oracle. We propose two algorithms termed Worcs-I and Worcs-II (Weak-Oracle Comparison- based Search), which provably locate the tar- get object in a number of comparisons close to the entropy of the target distribution. While Worcs-I provides better theoretical guarantees, Worcs-II is applicable to more technically challenging scenarios where the algorithm has limited access to the ranking dis- similarity between objects. A series of experiments validate the performance of our proposed algorithms.} }
Endnote
%0 Conference Paper %T Comparison Based Learning from Weak Oracles %A Ehsan Kazemi %A Lin Chen %A Sanjoy Dasgupta %A Amin Karbasi %B Proceedings of the Twenty-First International Conference on Artificial Intelligence and Statistics %C Proceedings of Machine Learning Research %D 2018 %E Amos Storkey %E Fernando Perez-Cruz %F pmlr-v84-kazemi18a %I PMLR %P 1849--1858 %U https://proceedings.mlr.press/v84/kazemi18a.html %V 84 %X There is increasing interest in learning algorithms that involve interaction between hu- man and machine. Comparison-based queries are among the most natural ways to get feed- back from humans. A challenge in designing comparison-based interactive learning algorithms is coping with noisy answers. The most common fix is to submit a query several times, but this is not applicable in many situations due to its prohibitive cost and due to the unrealistic assumption of independent noise in different repetitions of the same query. In this paper, we introduce a new weak oracle model, where a non-malicious user responds to a pairwise comparison query only when she is quite sure about the answer. This model is able to mimic the behavior of a human in noise-prone regions. We also consider the ap- plication of this weak oracle model to the problem of content search (a variant of the nearest neighbor search problem) through comparisons. More specifically, we aim at devising efficient algorithms to locate a target object in a database equipped with a dissimilarity metric via invocation of the weak comparison oracle. We propose two algorithms termed Worcs-I and Worcs-II (Weak-Oracle Comparison- based Search), which provably locate the tar- get object in a number of comparisons close to the entropy of the target distribution. While Worcs-I provides better theoretical guarantees, Worcs-II is applicable to more technically challenging scenarios where the algorithm has limited access to the ranking dis- similarity between objects. A series of experiments validate the performance of our proposed algorithms.
APA
Kazemi, E., Chen, L., Dasgupta, S. & Karbasi, A.. (2018). Comparison Based Learning from Weak Oracles. Proceedings of the Twenty-First International Conference on Artificial Intelligence and Statistics, in Proceedings of Machine Learning Research 84:1849-1858 Available from https://proceedings.mlr.press/v84/kazemi18a.html.

Related Material