Study shows projected neural additive models can act as universal approximators
A new arXiv paper argues that projected neural additive models, which trade some of the interpretability limits of standard NAMs, can serve as universal approximators, narrowing the expressivity gap with multi-layer perceptrons.
Read the full story at arXiv cs.LG ↗
Timeline · 1 report
- 2026-11-10 05:00 · arXiv cs.LG
Projected Neural Additive Models as Universal Approximators