Research map / Optimization / Nonconvex and smooth optimization
Boosting Frank-Wolfe by Chasing Gradients
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 |
|---|---|---|---|
| Chase the negative gradient using matching-pursuit steps rather than using only the vanilla Frank-Wolfe direction | Frank-Wolfe · Sparse signal recovery | function value per iteration | Helped |
| Chase the negative gradient using matching-pursuit steps rather than using only the vanilla Frank-Wolfe direction | Frank-Wolfe · Traffic assignment | function value per oracle call | Helped |
| Use line search instead of the short-step rule (ablation) | BoostFW · Sparsity-constrained logistic regression (Gisette) | function value per iteration | Helped |
| Use the short-step rule instead of line search (ablation) | BoostFW · Sparsity-constrained logistic regression (Gisette) | CPU time | Helped |
| Use the short-step rule instead of line search (ablation) | BoostFW · Collaborative filtering (MovieLens 100k) | function value per iteration | Helped |
| Use the line-search variant (ablation) | BoostFW · Collaborative filtering (MovieLens 100k) | CPU time | Helped |
| Use the short-step strategy instead of line search (ablation) | BoostFW · Sparse signal recovery | function value | Mixed |
| Set the pursuit-round cap to 1, reducing BoostFW to Frank-Wolfe (ablation) | BoostFW · Frank-Wolfe optimization | primal-gap convergence | Hurt |
| Apply gradient pursuit to DICG | DICG · Video co-localization (YouTube-Objects) | duality gap | Helped |
| Use DICG rather than BoostFW | Frank-Wolfe-style methods · Video co-localization (YouTube-Objects) | duality gap | Helped |
| Use line search instead of the short-step rule (ablation) | Away-Step Frank-Wolfe · Strongly convex optimization over polytopes | convergence rate | No clear effect |
Closest papers by meaning
- Beyond Short Steps in Frank-Wolfe Algorithms (ICLR 2026 · similarity 0.93)
- Revisiting Frank-Wolfe for Structured Nonconvex Optimization (NeurIPS 2025 · similarity 0.90)
- Linear Convergence of a Frank-Wolfe Type Algorithm over Trace-Norm Balls (NeurIPS 2017 · similarity 0.90)
- Sarah Frank-Wolfe: Methods for Constrained Optimization with Best Rates and Practical Features (ICML 2024 · similarity 0.89)
- Efficient Quadratic Corrections for Frank-Wolfe Algorithms (NeurIPS 2025 · similarity 0.88)
- Secant Line Search for Frank-Wolfe Algorithms (ICML 2025 · 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.