Effective Version Space Reduction for Convolutional Neural Networks
This work addresses inconsistency in active learning for neural networks, offering a principled approach to improve sampling efficiency, though it is incremental as it builds on existing version space reduction concepts.
The paper tackles the problem of sampling bias in active learning for convolutional neural networks by examining version space reduction, showing that diameter-based methods reduce version space more effectively and perform better than prior mass reduction and other baselines, with Gibbs vote disagreement matching the best query method on datasets like MNIST and SVHN.
In active learning, sampling bias could pose a serious inconsistency problem and hinder the algorithm from finding the optimal hypothesis. However, many methods for neural networks are hypothesis space agnostic and do not address this problem. We examine active learning with convolutional neural networks through the principled lens of version space reduction. We identify the connection between two approaches---prior mass reduction and diameter reduction---and propose a new diameter-based querying method---the minimum Gibbs-vote disagreement. By estimating version space diameter and bias, we illustrate how version space of neural networks evolves and examine the realizability assumption. With experiments on MNIST, Fashion-MNIST, SVHN and STL-10 datasets, we demonstrate that diameter reduction methods reduce the version space more effectively and perform better than prior mass reduction and other baselines, and that the Gibbs vote disagreement is on par with the best query method.