Official Implementation of "Generalization Error Analysis for Selective State-Space Models Through the Lens of Attention"
-
Updated
Nov 4, 2025 - Jupyter Notebook
Official Implementation of "Generalization Error Analysis for Selective State-Space Models Through the Lens of Attention"
Training Distribution Selection for Provable OOD Performance
🔥🔥🔥 This repository curates research on Weak-to-Strong Generalization across LLMs, multimodal learning, and beyond, focusing on how strong models learn from weak supervision and surpass their teachers. Stay tuned for the latest updates!
Generalization bounds for two-layer ReLU neural networks via scale-invariant complexity measures
Machine-checked Lean 4 / Mathlib formalisation of the universal spectral lower bound (Theorem 1) — formal-proof companion to structural-bounds-framework (zero sorry).
Out-of-distribution generalization bounds for neural image compression. Connects PAC-Bayes, transport theory, and rate-distortion.
Tight PAC-Bayes generalization bounds via numerical inversion of the binary KL divergence. Compares bisection and Newton's method for Maurer's PAC-Bayes-kl bound.
To associate your repository with the generalization-bounds topic, visit your repo's landing page and select "manage topics."