Abstract:Latent categorical variables are frequently found in deep learning architectures. They can model actions in discrete reinforcement-learning environments, represent categories in latent-variable models, or express relations in graph neural networks. Despite their widespread use, their discrete nature poses significant challenges to gradient-descent learning algorithms. While a substantial body of work has offered improved gradient estimation techniques, we take a complementary approach. Specifically, we: 1) revisit the ubiquitous $\textit{softmax}$ function and demonstrate its limitations from an information-geometric perspective; 2) replace the $\textit{softmax}$ with the $\textit{catnat}$ function, a function composed of a sequence of hierarchical binary splits; we prove that this choice offers significant advantages to gradient descent due to the resulting diagonal Fisher Information Matrix. A rich set of experiments - including graph structure learning, variational autoencoders, and reinforcement learning - empirically show that the proposed function improves the learning efficiency and yields models characterized by consistently higher test performance. $\textit{Catnat}$ is simple to implement and seamlessly integrates into existing codebases. Moreover, it remains compatible with standard training stabilization techniques and, as such, offers a better alternative to the $\textit{softmax}$ function.
| Subjects: | Machine Learning (cs.LG); Machine Learning (stat.ML) |
| Cite as: | arXiv:2509.24728 [cs.LG] |
| (or arXiv:2509.24728v2 [cs.LG] for this version) | |
| https://doi.org/10.48550/arXiv.2509.24728 arXiv-issued DOI via DataCite |
Submission history
From: Alessandro Manenti [view email]
[v1]
Mon, 29 Sep 2025 12:55:50 UTC (241 KB)
[v2]
Tue, 12 May 2026 20:11:36 UTC (1,351 KB)