Explanation-Based Attention for Semi-Supervised Deep Active Learning

Abstract

We introduce an attention mechanism to improve feature extraction for deep active learning (AL) in the semi-supervised setting. The proposed attention mechanism is based on recent methods to visually explain predictions made by DNNs. We apply the proposed explanation-based attention to MNIST and SVHN classification. The conducted experiments show accuracy improvements for the original and class-imbalanced datasets with the same number of training examples and faster long-tail convergence compared to uncertainty-based methods.

Cite

Text

Gudovskiy et al. "Explanation-Based Attention for Semi-Supervised Deep Active Learning." ICLR 2019 Workshops: LLD, 2019.

Markdown

[Gudovskiy et al. "Explanation-Based Attention for Semi-Supervised Deep Active Learning." ICLR 2019 Workshops: LLD, 2019.](https://mlanthology.org/iclrw/2019/gudovskiy2019iclrw-explanationbased/)

BibTeX

@inproceedings{gudovskiy2019iclrw-explanationbased,
  title     = {{Explanation-Based Attention for Semi-Supervised Deep Active Learning}},
  author    = {Gudovskiy, Denis and Hodgkinson, Alec and Yamaguchi, Takuya and Tsukizawa, Sotaro},
  booktitle = {ICLR 2019 Workshops: LLD},
  year      = {2019},
  url       = {https://mlanthology.org/iclrw/2019/gudovskiy2019iclrw-explanationbased/}
}