Research map / Optimization / Nonconvex and smooth optimization
Algorithmic Instabilities of Accelerated Gradient Descent
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's title only (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 |
|---|---|---|---|
| Apply NAG to general convex smooth objectives instead of restricting to quadratics | NAG · Convex smooth objectives | initialization stability: O(T\epsilon) -> \exp(\Theta(T))\epsilon | Hurt |
| Apply full-batch NAG to general convex smooth empirical risks instead of restricting to quadratics | Full-batch NAG · Convex smooth empirical risk | uniform stability: \Theta\brk{\ifrac{T^2}{n}} -> \ifrac{\exp(\Theta(T))}{n} | Hurt |
| Use NAG rather than GD in the general convex smooth setting | Convex smooth optimization · Convex smooth objectives | initialization stability: \Theta(\epsilon) -> \exp(\Theta(T))\epsilon | Hurt |
| Use full-batch NAG rather than GD in the general convex smooth setting | Full-batch optimization · Convex smooth empirical risk | uniform stability: \Theta\brk{\ifrac{T}{n}} -> \ifrac{\exp(\Theta(T))}{n} | Hurt |
Closest papers by meaning
- Breaking the Convergence Barrier: Optimization via Fixed-Time Convergent Flows (AAAI 2022 · similarity 0.88)
- Uniform Stability and Generalization Error of GD and SGD on Fixed-Point Parameters (NeurIPS 2026 · similarity 0.88)
- Directional Smoothness and Gradient Methods: Convergence and Adaptivity (NeurIPS 2024 · similarity 0.87)
- Convergence Rates for Gradient Descent on the Edge of Stability for Overparametrised Least Squares (NeurIPS 2025 · similarity 0.87)
- Generalization Analysis of Biased Stochastic Gradient Methods for Minimax Problems (NeurIPS 2026 · similarity 0.87)
- Nesterov acceleration despite very noisy gradients (NeurIPS 2024 · similarity 0.86)
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.