The Analysis of Deep Neural Networks by Information Theory: From Explainability to Generalization

Abstract

Despite their great success in many artificial intelligence tasks, deep neural networks (DNNs) still suffer from a few limitations, such as poor generalization behavior for out-of-distribution (OOD) data and the "black-box" nature. Information theory offers fresh insights to solve these challenges. In this short paper, we briefly review the recent developments in this area, and highlight our contributions.

Cite

Text

Yu. "The Analysis of Deep Neural Networks by Information Theory: From Explainability to Generalization." AAAI Conference on Artificial Intelligence, 2023. doi:10.1609/AAAI.V37I13.26829

Markdown

[Yu. "The Analysis of Deep Neural Networks by Information Theory: From Explainability to Generalization." AAAI Conference on Artificial Intelligence, 2023.](https://mlanthology.org/aaai/2023/yu2023aaai-analysis/) doi:10.1609/AAAI.V37I13.26829

BibTeX

@inproceedings{yu2023aaai-analysis,
  title     = {{The Analysis of Deep Neural Networks by Information Theory: From Explainability to Generalization}},
  author    = {Yu, Shujian},
  booktitle = {AAAI Conference on Artificial Intelligence},
  year      = {2023},
  pages     = {15462},
  doi       = {10.1609/AAAI.V37I13.26829},
  url       = {https://mlanthology.org/aaai/2023/yu2023aaai-analysis/}
}