Statistical active learning algorithms for noise tolerance and differential privacy

DOI10.1007/S00453-014-9954-9MaRDI QIDQ2345952zbMATH OpenOpenAlexFDO

Authors Maria-Florina Balcan, Vitaly Feldman

Publication date 21 May 2015

Published in Algorithmica (Search for Journal in Brave)

Full work available at URL https://arxiv.org/abs/1307.3102

zbMATH Keywords

differential privacyactive learningcomputationally efficient algorithmsrandom classification noisestatistical query framework

Mathematics Subject Classification ID

Learning and adaptive systems in artificial intelligence (68T05)

Abstract: We describe a framework for designing efficient active learning algorithms that are tolerant to random classification noise and are differentially-private. The framework is based on active learning algorithms that are statistical in the sense that they rely on estimates of expectations of functions of filtered random examples. It builds on the powerful statistical query framework of Kearns (1993). We show that any efficient active statistical learning algorithm can be automatically converted to an efficient active learning algorithm which is tolerant to random classification noise as well as other forms of "uncorrelated" noise. The complexity of the resulting algorithms has information-theoretically optimal quadratic dependence on $1/(1-2eta)$, where $eta$ is the noise rate. We show that commonly studied concept classes including thresholds, rectangles, and linear separators can be efficiently actively learned in our framework. These results combined with our generic conversion lead to the first computationally-efficient algorithms for actively learning some of these concept classes in the presence of random classification noise that provide exponential improvement in the dependence on the error $epsilon$ over their passive counterparts. In addition, we show that our algorithms can be automatically converted to efficient active differentially-private algorithms. This leads to the first differentially-private active learning algorithms with exponential label savings over the passive case.

Recommendations

• Efficient noise-tolerant learning from statistical queries
• What can we learn privately?
• Active Learning in the Non-realizable Case
• Agnostic active learning
• On the efficiency of noise-tolerant PAC algorithms derived from statistical queries

Cites work

• scientific article; zbMATH DE number 1332320 (Why is no real title available?)
• 10.1162/153244302760200669
• A complete characterization of statistical query learning with applications to evolvability
• A polynomial-time algorithm for learning noisy linear threshold functions
• A simple polynomial-time rescaling algorithm for solving linear programs
• A theory of the learnable
• Agnostic active learning
• Analysis of perceptron-based active learning
• Efficient active learning of halfspaces: an aggressive approach
• Efficient noise-tolerant learning from statistical queries
• Learnability and the Vapnik-Chervonenkis dimension
• Learning noisy linear classifiers via adaptive and selective sampling
• Margin Based Active Learning
• Minimax Bounds for Active Learning
• Rademacher complexities and bounding the excess risk in active learning
• Selective sampling and active learning from single and multiple teachers
• Selective sampling using the query by committee algorithm
• Smoothness, disagreement coefficient, and the label complexity of agnostic active learning
• Specification and simulation of statistical query algorithms for efficiency and noise tolerance
• Statistical algorithms and a lower bound for detecting planted cliques
• Teaching Dimension and the Complexity of Active Learning
• The geometry of logconcave functions and sampling algorithms
• The true sample complexity of active learning
• Theory of Classification: a Survey of Some Recent Advances
• Theory of Cryptography
• Theory of Disagreement-Based Active Learning
• Weakly learning DNF and characterizing statistical query learning using Fourier analysis
• What can we learn privately?

Cited in

(5)

• On the complexity of random satisfiability problems with planted solutions
• The Limitations of Optimization from Samples
• Active learning for noisy oracle via density power divergence
• On the Power of Learning from k-Wise Queries
• Active learning in heteroscedastic noise

Describes a project that uses

Uses Software

• SuLQ

This page was built for publication: Statistical active learning algorithms for noise tolerance and differential privacy

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2345952)