Research map / Bandits and online learning / Online convex optimization and games
Beyond Bandit Feedback in Online Multiclass Classification
Dirk van der Hoeven, Federico Fusco, Nicolò Cesa-Bianchi
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 GapHin (hinge loss) instead of GapLog or GapSmH (ablation) | Gappletron · synthetic bandit setting | error rate stability | Hurt |
| use logistic loss (GapLog) instead of hinge or smooth hinge (ablation) | Gappletron · synthetic multiclass spam filtering, K > 6 | mistakes | Hurt |
| use logistic loss (GapLog) instead of hinge or smooth hinge (ablation) | Gappletron · synthetic bandit setting | range of error rate over repetitions | Improved |
| increase the exploration rate (ablation) | Gappletron · synthetic data, all settings | stability of error rate | Improved |
| use Gappletron instead of PNewtron, SOBAdiag and importance weighted Banditron | online multiclass classifier · synthetic bandit setting | error rate | Improved |
| restrict label requests to the revealing action (label efficient feedback graph, rho=1) instead of bandit feedback (rho=K) | Gappletron surrogate regret bound · theory | surrogate regret bound | Improved |
Closest papers by meaning
- Bandit and Delayed Feedback in Online Structured Prediction (NeurIPS 2025 · similarity 0.90)
- Bandit-Feedback Online Multiclass Classification: Variants and Tradeoffs (NeurIPS 2024 · similarity 0.90)
- Efficient Contextual Bandits with Uninformed Feedback Graphs (ICML 2024 · similarity 0.88)
- Fast Rates for Bandit PAC Multiclass Classification (NeurIPS 2024 · similarity 0.87)
- From Contextual Combinatorial Semi-Bandits to Bandit List Classification: Improved Sample Complexity with Sparse Rewards (NeurIPS 2025 · similarity 0.87)
- A Universal Growth Rate for Learning with Smooth Surrogate Losses (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.
