Research map / Optimization / Nonconvex and smooth optimization
Faster Online Learning of Optimal Threshold for Consistent F-measure Optimization
Xiaoxuan Zhang, Mingrui Liu, Xun Zhou, Tianbao Yang
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 |
|---|---|---|---|
| use FOFO threshold learning instead of the OFO heuristic threshold update | FOFO · covtype 2 vs o and 1 vs o (balanced) | online F-score | No clear effect |
| use FOFO instead of OFO on highly imbalanced data | FOFO · covtype 6 vs o, 5 vs o | online F-score | Improved |
| increase class imbalance of the dataset | FOFO vs baselines · covtype (p from 48.76% to 1.63%) | online F-score improvement | Improved |
| use LR with fixed 0.5 threshold instead of threshold learning | LR · covtype 6 vs o, 5 vs o | online F-score | Hurt |
| use FOFO instead of STAMP and OMCSL | FOFO · 25 binary tasks from 7 LIBSVM datasets | online F-score | Improved |
| use FOFO instead of OFO and others | FOFO · covtype, webspam, a9a, ijcnn1, w8a, sensorless, protein | F-score vs running time | Improved |
| learn threshold by strongly convex stochastic optimization, replacing the heuristic update | OFO · Theory, threshold convergence | convergence rate | Improved |
| FOFO versus OMCSL regarding memory and compute | FOFO · Table 1 | memory cost | Improved |
Closest papers by meaning
- Extreme F-measure Maximization using Sparse Probability Estimates (ICML 2016 · similarity 0.86)
- Convex Calibrated Surrogates for the Multi-Label F-Measure (ICML 2020 · similarity 0.86)
- Fast Rates of ERM and Stochastic Approximation: Adaptive to Error Bound Conditions (NeurIPS 2018 · similarity 0.85)
- Sarah Frank-Wolfe: Methods for Constrained Optimization with Best Rates and Practical Features (ICML 2024 · similarity 0.84)
- A General Online Algorithm for Optimizing Complex Performance Metrics (ICML 2024 · similarity 0.84)
- Stochastic Zeroth-Order Optimization under Strongly Convexity and Lipschitz Hessian: Minimax Sample Complexity (NeurIPS 2024 · similarity 0.84)
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.
