A Unified Batch Selection Policy for Active Metric Learning

Active metric learning is the problem of incrementally selecting high-utility batches of training data (typically, ordered triplets) to annotate, in order to progressively improve a learned model of a metric over some input domain as rapidly as possible. Standard approaches, which independently asse...

Full description

Saved in:
Bibliographic Details
Published inMachine Learning and Knowledge Discovery in Databases. Research Track Vol. 12976; pp. 599 - 616
Main Authors Priyadarshini, K., Chaudhuri, Siddhartha, Borkar, Vivek, Chaudhuri, Subhasis
Format Book Chapter
LanguageEnglish
Published Switzerland Springer International Publishing AG 2021
Springer International Publishing
SeriesLecture Notes in Computer Science
Subjects
Batch active learning
Maximum Entropy Principle
Perceptual metric
Submodular optimization
Online AccessGet full text
ISBN3030865193
9783030865191
ISSN0302-9743
1611-3349
DOI10.1007/978-3-030-86520-7_37

Cover

More Information
Summary:Active metric learning is the problem of incrementally selecting high-utility batches of training data (typically, ordered triplets) to annotate, in order to progressively improve a learned model of a metric over some input domain as rapidly as possible. Standard approaches, which independently assess the informativeness of each triplet in a batch, are susceptible to highly correlated batches with many redundant triplets and hence low overall utility. While a recent work [20] proposes batch-decorrelation strategies for metric learning, they rely on ad hoc heuristics to estimate the correlation between two triplets at a time. We present a novel batch active metric learning method that leverages the Maximum Entropy Principle to learn the least biased estimate of triplet distribution for a given set of prior constraints. To avoid redundancy between triplets, our method collectively selects batches with maximum joint entropy, which simultaneously captures both informativeness and diversity. We take advantage of the submodularity of the joint entropy function to construct a tractable solution using an efficient greedy algorithm based on Gram-Schmidt orthogonalization that is provably 1-1e $$\left( 1 - \frac{1}{e} \right) $$ -optimal. Our approach is the first batch active metric learning method to define a unified score that balances informativeness and diversity for an entire batch of triplets. Experiments with several real-world datasets demonstrate that our algorithm is robust, generalizes well to different applications and input modalities, and consistently outperforms the state-of-the-art.
Bibliography:Electronic supplementary materialThe online version of this chapter (https://doi.org/10.1007/978-3-030-86520-7_37) contains supplementary material, which is available to authorized users.
Original Abstract: Active metric learning is the problem of incrementally selecting high-utility batches of training data (typically, ordered triplets) to annotate, in order to progressively improve a learned model of a metric over some input domain as rapidly as possible. Standard approaches, which independently assess the informativeness of each triplet in a batch, are susceptible to highly correlated batches with many redundant triplets and hence low overall utility. While a recent work [20] proposes batch-decorrelation strategies for metric learning, they rely on ad hoc heuristics to estimate the correlation between two triplets at a time. We present a novel batch active metric learning method that leverages the Maximum Entropy Principle to learn the least biased estimate of triplet distribution for a given set of prior constraints. To avoid redundancy between triplets, our method collectively selects batches with maximum joint entropy, which simultaneously captures both informativeness and diversity. We take advantage of the submodularity of the joint entropy function to construct a tractable solution using an efficient greedy algorithm based on Gram-Schmidt orthogonalization that is provably 1-1e\documentclass[12pt]{minimal} \usepackage{amsmath} \usepackage{wasysym} \usepackage{amsfonts} \usepackage{amssymb} \usepackage{amsbsy} \usepackage{mathrsfs} \usepackage{upgreek} \setlength{\oddsidemargin}{-69pt} \begin{document}$$\left( 1 - \frac{1}{e} \right) $$\end{document}-optimal. Our approach is the first batch active metric learning method to define a unified score that balances informativeness and diversity for an entire batch of triplets. Experiments with several real-world datasets demonstrate that our algorithm is robust, generalizes well to different applications and input modalities, and consistently outperforms the state-of-the-art.
ISBN:3030865193
9783030865191
ISSN:0302-9743
1611-3349
DOI:10.1007/978-3-030-86520-7_37