A Self-Generated Floor on Adam's epsilon: Evidence from Power-Law Matrix Sensing
cs.LG
Submitted: 2026-11-01
Updated: 2026-11-01
Terminology
Sources
- Adaptive Gradient Methods at the Edge of Stability
- Scaling Exponents Across Parameterizations and Optimizers
- Adam: A Method for Stochastic Optimization
- Heavy-Tailed Class Imbalance and Why Adam Outperforms Gradient Descent on Language Models
- The Precision Wall: An Optimizer-Dependent Boundary for Optimization under Power-Law Measurement Spectra
Related papers
- A Physics-Regulated Neural Framework for Learning 3D Grain Growth Dynamics
- FlexMoE: Learning One-for-All Multi-Budget Expert Pruning for Flexible MoE Deployment
- OptiMind: Improving Small Language Models for Optimization Modeling through Formulation Guidance and Multi-Turn Learning
- Staying in the Sweet Spot: Boost Reasoning Evolution via Responsive Rollout Difficulty Manipulation
- Examining the Impact of Data and Training Methods on Emergent and Subliminal Misalignment
- Rethinking Black-Box On-Policy Distillation: A Rubric-based Analysis