Edge of Stability Selectively Shapes Learning Across the Data Distribution
cs.LG, stat.ML
Submitted: 2026-06-02
Updated: 2026-11-12
Code: https://github.com/arseniqum/edge-of-stochastic-stability
Terminology
Sources
- Understanding deep learning requires rethinking generalization
- Three Factors Influencing Minima in SGD
- A Walk with SGD
- On the Relation Between the Sharpest Directions of DNN Loss and the SGD Step Length
- The Break-Even Point on Optimization Trajectories of Deep Neural Networks
- Edge of Stochastic Stability: Revisiting the Edge of Stability for SGD
- Momentum Further Constrains Sharpness at the Edge of Stochastic Stability
- Non-Euclidean Gradient Descent Operates at the Edge of Stability
- A Scalable Measure of Loss Landscape Curvature for Analyzing the Training Dynamics of LLMs
- Does Weight Decay Enhance Training Stability?
- Too Sharp, Too Sure: When Calibration Follows Curvature
- Deep Learning on a Data Diet: Finding Important Examples Early in Training
- The large learning rate phase of deep learning: the catapult mechanism
- Towards Understanding Generalization of Deep Learning: Perspective of Loss Landscapes
- Adam: A Method for Stochastic Optimization
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