Research map / Optimization / Nonconvex and smooth optimization
Asynchronous Coordinate Descent under More Realistic Assumptions
Tao Sun, Robert Hannah, Wotao Yin
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 |
|---|---|---|---|
| Allow delays to depend on the updated block | Async-BCD · Bounded delays, deterministic blocks, nonconvex | gradient norm convergence | Improved |
| Allow delays to depend on the updated block | Async-BCD · Bounded delays, stochastic blocks, nonconvex | expected gradient norm convergence | Improved |
| Allow unbounded stochastic delays with sufficiently fast-decaying tails | Async-BCD · Stochastic unbounded delays, deterministic blocks | expected gradient norm convergence | Improved |
| Allow unbounded stochastic delays with sufficiently fast-decaying tails | Async-BCD · Stochastic unbounded delays, stochastic blocks | expected gradient norm convergence | Improved |
| Use delay-adaptive step sizes for deterministic unbounded delays | Async-BCD · Deterministic unbounded delays | update norm | Improved |
| Model delays directly in a Lyapunov function instead of treating as noise | Async-BCD · Bounded delays | running-best update norm | Improved |
| Use deterministic block choices instead of random block selection | Async-BCD · Bounded delays, nonconvex | gradient norm convergence | Improved |
| Require a finite delay bound for deterministic unbounded delays | Async-BCD · Deterministic unbounded delays | gradient norm on Q_T | Mixed |
Closest papers by meaning
- Asynchronous Iterations in Optimization: New Sequence Results and Sharper Algorithmic Guarantees (JMLR 2023 · similarity 0.93)
- A2BCD: Asynchronous Acceleration with Optimal Complexity (ICLR 2019 · similarity 0.92)
- Distributed Asynchronous Optimization with Unbounded Delays: How Slow Can You Go? (ICML 2018 · similarity 0.89)
- Delay-agnostic Asynchronous Coordinate Update Algorithm (ICML 2023 · similarity 0.89)
- Do We Need Asynchronous SGD? On the Near-Optimality of Synchronous Methods (NeurIPS 2026 · similarity 0.88)
- On the Optimal Time Complexities in Decentralized Stochastic Asynchronous Optimization (NeurIPS 2024 · similarity 0.87)
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.
