Research map / Optimization / Nonconvex and smooth optimization
Algorithms and matching lower bounds for approximately-convex optimization
Andrej Risteski, Yuanzhi Li
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 |
|---|---|---|---|
| Information-theoretic lower bound construction: hide a random linear function inside a non-linear, non-smooth function with approximation error Delta = max(eps^2/sqrt(d), eps/d) (up to polylog) | approximately-convex zero-order optimization over convex bodies · unit-diameter convex body in R^d, 1-Lipschitz convex f | queries needed to find eps-optimal point | Improved |
| Algorithm using a robust zeroth-order gradient estimator (Stokes-formula smoothing) with extension of the function outside K via distance penalty | approximately-convex optimization over well-conditioned convex bodies · Delta <= max(eps^2/(mu sqrt(d)), eps/d) with mu-well-conditioned body | polynomial runtime in d and 1/eps | Improved |
| Prior local-search method of DKS14 limited to Delta = O(eps/sqrt(d)) for smooth functions; new method extends to Delta up to eps^2/sqrt(d) | approximately-convex optimization · theoretical comparison with prior work | maximum tolerable approximation error Delta | Improved |
Closest papers by meaning
- Near-Optimal Lower Bounds For Convex Optimization For All Orders of Smoothness (NeurIPS 2021 · similarity 0.88)
- Tight Complexity Bounds for Optimizing Composite Objectives (NeurIPS 2016 · similarity 0.87)
- On Lower and Upper Bounds in Smooth and Strongly Convex Optimization (JMLR 2016 · similarity 0.87)
- Information Complexity of Stochastic Convex Optimization: Applications to Generalization, Memorization, and Tracing (ICML 2024 · similarity 0.85)
- Faster Accelerated First-order Methods for Convex Optimization with Strongly Convex Function Constraints (NeurIPS 2024 · similarity 0.85)
- Single-Loop Stochastic Algorithms for Difference of Max-Structured Weakly Convex Functions (NeurIPS 2024 · similarity 0.85)
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.
