Efficient and Parsimonious Agnostic Active Learning

Abstract

We develop a new active learning algorithm for the streaming setting satisfying three important properties: 1) It provably works for any classifier representation and classification problem including those with severe noise. 2) It is efficiently implementable with an ERM oracle. 3) It is more aggressive than all previous approaches satisfying 1 and 2. To do this we create an algorithm based on a newly defined optimization problem and analyze it. We also conduct the first experimental analysis of all efficient agnostic active learning algorithms, evaluating their strengths and weaknesses in different settings.

References in corpus (3)

• Agnostic Active Learning Without Constraints
• Online Importance Weight Aware Updates
• Beyond Disagreement-based Agnostic Active Learning

Cited by in corpus (11)

• Deep Active Learning over the Long Tail
• A Contextual Bandit Bake-off
• Relaxed Softmax for learning from Positive and Unlabeled data
• Search Improves Label for Active Learning
• Improved Algorithms for Agnostic Pool-based Active Classification
• Adaptive Region-Based Active Learning
• Disagreement-Based Combinatorial Pure Exploration: Sample Complexity Bounds and an Efficient Algorithm
• On the Power of Localized Perceptron for Label-Optimal Learning of Halfspaces with Adversarial Noise
• The Label Complexity of Active Learning from Observational Data
• Learning to Actively Learn: A Robust Approach
• Corruption Robust Active Learning