Smooth InfoMax - Towards Easier Post-Hoc Interpretability

Abstract

We introduce Smooth InfoMax (SIM), a self-supervised representation learning method that incorporates interpretability constraints into the latent representations at different depths of the network. Based on $\beta $ β -VAEs, SIM’s architecture consists of probabilistic modules optimized locally with the InfoNCE loss to produce Gaussian-distributed representations regularized toward the standard normal distribution. This creates smooth, well-defined, and better-disentangled latent spaces, enabling easier post-hoc analysis. Evaluated on speech data, SIM preserves the large-scale training benefits of Greedy InfoMax while improving the effectiveness of post-hoc interpretability methods across layers. Our code is available via GitHub .

Cite

Text

Denoodt et al. "Smooth InfoMax - Towards Easier Post-Hoc Interpretability." European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, 2025. doi:10.1007/978-3-032-06066-2_30

Markdown

[Denoodt et al. "Smooth InfoMax - Towards Easier Post-Hoc Interpretability." European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases, 2025.](https://mlanthology.org/ecmlpkdd/2025/denoodt2025ecmlpkdd-smooth/) doi:10.1007/978-3-032-06066-2_30

BibTeX

@inproceedings{denoodt2025ecmlpkdd-smooth,
  title     = {{Smooth InfoMax - Towards Easier Post-Hoc Interpretability}},
  author    = {Denoodt, Fabian and de Boer, Bart and Oramas, José},
  booktitle = {European Conference on Machine Learning and Principles and Practice of Knowledge Discovery in Databases},
  year      = {2025},
  pages     = {512-527},
  doi       = {10.1007/978-3-032-06066-2_30},
  url       = {https://mlanthology.org/ecmlpkdd/2025/denoodt2025ecmlpkdd-smooth/}
}