Rare Gate Disagreements Can Limit Plasticity: When Gradient Flow Mispredicts Finite-Batch SGD
cs.LG, math.OC, stat.ML
Submitted: 2026-10-08
Updated: 2026-10-08
Terminology
Sources
- Directions of Curvature as an Explanation for Loss of Plasticity
- Gradient Descent Quantizes ReLU Network Features
- Predicting Plasticity in Deep Continual Learning: A Theoretical Perspective
- Type-II Saddles and Probabilistic Stability of Stochastic Gradient Descent
Related papers
- Polynomial-Augmented Neural Networks (PANNs) with Weak Orthogonality Constraints for Enhanced Function and PDE Approximation
- AIRL-S: Unifying Reinforcement Learning and Search-Based Test-Time Scaling via Adversarial Inverse Reinforcement Learning
- Transformers as Bayesian In-Context Experimenters: Smoothness-Adaptive Efficient ATE Estimation
- Convergence issues in Relational Concept Analysis based on AOC-posets
- Beliefs Beyond Posteriors: Local-Consistency Optimisation for Bayesian Neural Networks
- Understanding Diffusion Models via Ratio-Based Function Approximation with SignReLU Networks