Towards Understanding Momentum Acceleration in River-Valley Loss Landscape
cs.LG, math.OC
Submitted: 2026-09-25
Updated: 2026-09-25
Terminology
Sources
- Momentum Further Constrains Sharpness at the Edge of Stochastic Stability
- Gradient Descent on Neural Networks Typically Occurs at the Edge of Stability
- Gradient descent with adaptive stepsize converges (nearly) linearly under fourth-order growth
- TinyStories: How Small Can Language Models Be and Still Speak Coherent English?
- When and Why Momentum Accelerates SGD:An Empirical Study
- The Pile: An 800GB Dataset of Diverse Text for Language Modeling
- Accurate, Large Minibatch SGD: Training ImageNet in 1 Hour
- Polyak's Heavy Ball Method Achieves Accelerated Local Rate of Convergence under Polyak-Lojasiewicz Inequality
- Adam: A Method for Stochastic Optimization
- Beyond the Quadratic Approximation: the Multiscale Structure of Neural Network Loss Landscapes
- Gradient Descent on Logistic Regression with Non-Separable Data and Large Step Sizes
- The AdEMAMix Optimizer: Better, Faster, Older
- Accelerated Over-Relaxation Heavy-Ball Method: Achieving Global Accelerated Convergence with Broad Generalization
- A Walk with SGD
- The Implicit Bias of Heterogeneity towards Invariance: A Study of Multi-Environment Matrix Sensing
- Minimax Optimal Convergence of Gradient Descent in Logistic Regression via Large and Adaptive Stepsizes
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