LGMLDec 15, 2019

Disentanglement based Active Learning

arXiv:1912.07018v2
Originality Incremental advance
AI Analysis

This work addresses the labeling budget problem in active learning for image classification, offering an incremental improvement over prior GAN-based approaches.

The paper tackles the problem of reducing human labeling costs in active learning by proposing Disentanglement based Active Learning (DAL), which uses self-supervision and disentanglement to automatically label most datapoints, achieving better performance on three benchmark image classification datasets compared to existing GAN-based methods.

We propose Disentanglement based Active Learning (DAL), a new active learning technique based on self-supervision which leverages the concept of disentanglement. Instead of requesting labels from human oracle, our method automatically labels the majority of the datapoints, thus drastically reducing the human labeling budget in Generative Adversarial Net (GAN) based active learning approaches. The proposed method uses Information Maximizing Generative Adversarial Nets (InfoGAN) to learn disentangled class category representations. Disagreement between active learner predictions and InfoGAN labels decides if the datapoints need to be human-labeled. We also introduce a label correction mechanism that aims to filter out label noise that occurs due to automatic labeling. Results on three benchmark datasets for the image classification task demonstrate that our method achieves better performance compared to existing GAN-based active learning approaches.

Code Implementations1 repo
Foundations

The foundational work for this paper's niche, ranked by how specifically the neighbourhood builds on it — not by global fame.

Your Notes