Convergence of gradient descent for deep neural networks
cs.LG, cs.NE, math.OC, math.PR, stat.ML
Submitted: 2022-03-30
Updated: 2026-09-21
Comments: 31 pages, 3 figures. Minor changes in this revision. To appear in J. Mach. Learn. Res
License: http://arxiv.org/licenses/nonexclusive-distrib/1.0/
Terminology
Sources
- Lower Bounds for Non-Convex Stochastic Optimization
- A Convergence Analysis of Gradient Descent for Deep Linear Neural Networks
- Recent Theoretical Advances in Non-Convex Optimization
- Gradient Descent Provably Optimizes Over-parameterized Neural Networks
- Qualitatively characterizing neural network optimization problems
- Identity Matters in Deep Learning
- On the existence of global minima and convergence analyses for gradient descent methods in the training of deep neural networks
- Gradient descent aligns the layers of deep linear networks
- An overview of gradient descent optimization algorithms
- Optimization for deep learning: theory and algorithms
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