Bayesian Active Learning by Disagreements: A Geometric Perspective

2021/05/06 by Xiaofeng Cao, Ivor W. Tsang, Cao, Xiaofeng +1
Computer Science · #Machine Learning and Algorithms #Computability, Logic, AI Algorithms #Algorithms and Data Compression

paper · pdf · doi:10.48550/arxiv.2105.02543

Abstract

We present geometric Bayesian active learning by disagreements (GBALD), a framework that performs BALD on its core-set construction interacting with model uncertainty estimation. Technically, GBALD constructs core-set on ellipsoid, not typical sphere, preventing low-representative elements from spherical boundaries. The improvements are twofold: 1) relieve uninformative prior and 2) reduce redundant estimations. Theoretically, geodesic search with ellipsoid can derive tighter lower bound on error and easier to achieve zero error than with sphere. Experiments show that GBALD has slight perturbations to noisy and repeated samples, and outperforms BALD, BatchBALD and other existing deep active learning approaches.

Citations

• The Efficiency of Some Nonparametric Competitors of the t-Test
• A Derivation of n-Dimensional Spherical Coordinates

Related

• Deep Bayesian Active Learning for Natural Language Processing: Results\n of a Large-Scale Empirical Study
• A Divide-and-Conquer Approach to Geometric Sampling for Active Learning
• Deep Bayesian Active Learning, A Brief Survey on Recent Advances
• Near Optimal Bayesian Active Learning for Decision Making
• Bayesian active learning for production, a systematic study and a reusable library