Research map / Optimization / Nonconvex and smooth optimization
Classic but Everlasting: Traditional Gradient-Based Algorithms Converge Fast Even in Time-Varying Multi-Player Games
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's abstract only. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Shows that traditional gradient-based algorithms (e.g., gradient descent/optimistic variants) achieve fast last-iterate convergence in time-varying multi-player games, extending beyond time-invariant two-player bilinear zero-sum settings.
Problem
Last-iterate convergence of gradient-based learning algorithms is known mainly for time-invariant games; prior results for time-varying games are limited to two-player bilinear time-varying zero-sum games.
Approach
Analyzes traditional gradient-based algorithms in multi-player games whose payoffs vary over time, showing fast last-iterate convergence beyond the two-player bilinear zero-sum setting. Details beyond the abstract are not available.
Key results
- Traditional gradient-based algorithms are shown to converge fast in the last iterate in time-varying multi-player games (per the title; quantitative rates not stated in the abstract provided).
Limitations
- Full text unavailable; only the abstract's claims are captured.
Closest papers by meaning
- Fast Last-Iterate Convergence of Learning in Games Requires Forgetful Algorithms (NeurIPS 2024 · similarity 0.93)
- Last-Iterate Convergence Properties of Regret-Matching Algorithms in Games (ICLR 2025 · similarity 0.91)
- From Average-Iterate to Last-Iterate Convergence in Games: A Reduction and Its Applications (NeurIPS 2025 · similarity 0.91)
- Convergence of $\text{log}(1/\epsilon)$ for Gradient-Based Algorithms in Zero-Sum Games without the Condition Number: A Smoothed Analysis (NeurIPS 2024 · similarity 0.91)
- Near-Optimal Last-Iterate Convergence for Zero-Sum Games with Bandit Feedback and Opponent Actions (NeurIPS 2026 · similarity 0.91)
- When and Why is Optimistic Multiplicative Weights Slow? The Geometry of Energy Dissipation (NeurIPS 2026 · similarity 0.91)
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.