Toward optimal probabilistic active learning using a Bayesian approach

DOI10.1007/S10994-021-05986-9MaRDI QIDQ2051315zbMATH OpenOpenAlexFDO

Authors Daniel Kottke, Marek Herde, Christoph Sandrock, Denis Huseljic, Georg Krempl, Bernhard Sick

Publication date 24 November 2021

Published in Machine Learning (Search for Journal in Brave)

Copyright license Creative Commons Attribution 4.0 International

Full work available at URL https://arxiv.org/abs/2006.01732

zbMATH Keywords

active learningclassificationprobabilistic active learning

Mathematics Subject Classification ID

Classification and discrimination; cluster analysis (statistical aspects) (62H30) Learning and adaptive systems in artificial intelligence (68T05)

Abstract: Gathering labeled data to train well-performing machine learning models is one of the critical challenges in many applications. Active learning aims at reducing the labeling costs by an efficient and effective allocation of costly labeling resources. In this article, we propose a decision-theoretic selection strategy that (1) directly optimizes the gain in misclassification error, and (2) uses a Bayesian approach by introducing a conjugate prior distribution to determine the class posterior to deal with uncertainties. By reformulating existing selection strategies within our proposed model, we can explain which aspects are not covered in current state-of-the-art and why this leads to the superior performance of our approach. Extensive experiments on a large variety of datasets and different kernels validate our claims.

Review text

Active learning methods select the next input to be learned in order to make the learning more efficient under the supervised learning, that is, the learning of input-output relationships. This paper considers the active learning framework in a classification setting. Using the Dirichlet conjugate prior for the categorical model, the authors define the estimated risk difference and propose a new active learning strategy, expected probabilistic gain for active learning (xPAL). With some modifications, xPAL reduces to existing active learning methods, expected error reduction (EER), probabilistic active learning (PAL), and uncertainty sampling (US). The authors show these relationships theoretically and compare the performance of xPAL with the existing methods numerically. Although little is discussed on the connections to active learning frameworks based on statistical experimental design, xPAL provides a unifying framework for the recent active learning methods.

Recommendations

• Optimised probabilistic active learning (OPAL). For fast, non-myopic, cost sensitive active classification
• Active learning with adaptive regularization
• Active learning
• Active learning for cost-sensitive classification
• Two faces of active learning

Cites work

• Active learning
• Algorithm runtime prediction: methods \& evaluation
• Evaluating learning algorithms. A classification perspective
• scientific article; zbMATH DE number 823069 (Why is no real title available?)
• Optimised probabilistic active learning (OPAL). For fast, non-myopic, cost sensitive active classification
• Pattern recognition and machine learning.
• Semi-supervised active learning for support vector machines: a novel approach that exploits structure information in data

Cited in

(23)

• Near-optimal Bayesian active learning with correlated and noisy tests
• Evidence-based uncertainty sampling for active learning
• Statistical asymptotic theory of active learning
• A batch ensemble approach to active learning with model selection
• A step-by-step tutorial on active inference and its application to empirical data
• Stream-based active learning for sliding windows under the influence of verification latency
• Active learning for hierarchical multi-label classification
• Active learning using transductive sparse Bayesian regression
• Nuclear discrepancy for single-shot batch active learning
• Two optimal strategies for active learning of causal models from interventional data
• A Lyapunov approach for stable reinforcement learning
• Active classification using belief functions and information gain maximization
• Active learning
• Active learning of Bayesian linear models with high-dimensional binary features by parameter confidence-region estimation
• Active learning with a misspecified prior
• Chance-Constrained Active Inference
• From active to proactive learning methods
• Plug-in approach to active learning
• Cost-Minimising Strategies for Data Labelling: Optimal Stopping and Active Learning
• A Bayesian reinforcement learning approach in Markov games for computing near-optimal policies
• Poisson Reweighted Laplacian Uncertainty Sampling for Graph-Based Active Learning
• Hessian optimal design for image retrieval
• Optimised probabilistic active learning (OPAL). For fast, non-myopic, cost sensitive active classification

Describes a project that uses

Uses Software

• PRMLT
• scikit-activeml
• OpenML

This page was built for publication: Toward optimal probabilistic active learning using a Bayesian approach

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q2051315)