Deep Ensemble Bayesian Active Learning : Addressing the Mode Collapse issue in Monte Carlo dropout via Ensembles

Abstract

In image classification tasks, the ability of deep CNNs to deal with complex image data has proven to be unrivalled. However, they require large amounts of labeled training data to reach their full potential. In specialised domains such as healthcare, labeled data can be difficult and expensive to obtain. Active Learning aims to alleviate this problem, by reducing the amount of labelled data needed for a specific task while delivering satisfactory performance. We propose DEBAL, a new active learning strategy designed for deep neural networks. This method improves upon the current state-of-the-art deep Bayesian active learning method, which suffers from the mode collapse problem. We correct for this deficiency by making use of the expressive power and statistical properties of model ensembles. Our proposed method manages to capture superior data uncertainty, which translates into improved classification performance. We demonstrate empirically that our ensemble method yields faster convergence of CNNs trained on the MNIST and CIFAR-10 datasets.

arXiv commentICLR under review

References in corpus (10)

• Less is more: sampling chemical space with active learning
• Deep Bayesian Active Learning with Image Data
• VEEGAN: Reducing Mode Collapse in GANs using Implicit Variational Learning
• Bayesian Active Learning for Classification and Preference Learning
• Understanding Measures of Uncertainty for Adversarial Example Detection
• Adversarial Active Learning for Deep Networks: a Margin Based Approach
• Variational Bayesian Inference with Stochastic Search
• Deep Active Learning for Named Entity Recognition
• Active and Continuous Exploration with Deep Neural Networks and Expected Model Output Changes
• Active Learning for Visual Question Answering: An Empirical Study

Cited by in corpus (6)

• A Review of Uncertainty Quantification in Deep Learning: Techniques, Applications and Challenges
• ScatterShot: Interactive In-context Example Curation for Text Transformation
• Deeper Connections between Neural Networks and Gaussian Processes Speed-up Active Learning
• Bayesian Estimate of Mean Proper Scores for Diversity-Enhanced Active Learning
• Neural Active Learning with Performance Guarantees
• Knowledge-driven Active Learning