Research map / Optimization / Implicit bias and generalization
Continuous Time Analysis of Momentum Methods
Nikola B. Kovachki, Andrew M. Stuart
Read the paperCheck your idea against 137,217 papers
AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Measured effects reported in the paper
| What was changed | Where | Measured | Effect |
|---|---|---|---|
| decrease the learning rate h (ablation) | HB and NAG · MNIST autoencoder | final training error gap between methods | Improved |
| use NAG instead of HB or gradient flow at large learning rate | MNIST autoencoder · MNIST autoencoder | final training error | Improved |
| use momentum lambda=0.9 with h=1 (effective learning rate 10) (ablation) | gradient flow discretisation (forward Euler) · MNIST autoencoder | trainability | Hurt |
| choose the learning-rate-dependent momentum lambda=(1-sqrt(mu h))/(1+sqrt(mu h)) with mu=1 (ablation) | HB-mu and NAG-mu · MNIST autoencoder | convergence rate to limiting ODE: around 1 vs around 0.5 | No clear effect |
Closest papers by meaning
- Losing Momentum in Continuous-time Stochastic Optimisation (JMLR 2025 · similarity 0.93)
- The Marginal Value of Momentum for Small Learning Rate SGD (ICLR 2024 · similarity 0.90)
- On the Performance Analysis of Momentum Method: A Frequency Domain Perspective (ICLR 2025 · similarity 0.90)
- Heavy-Ball Momentum Method in Continuous Time and Discretization Error Analysis (NeurIPS 2025 · similarity 0.90)
- Perfect Parallelization in Mini-Batch SGD with Classical Momentum Acceleration (NeurIPS 2026 · similarity 0.90)
- Gradient correlation is a key ingredient to accelerate SGD with momentum (ICLR 2025 · similarity 0.89)
Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.
