Research map / Bandits and online learning / Online convex optimization and games
Improved Algorithms for Online Classification with Surrogate Losses
Read the paperCheck your idea against 108,967 papers
AI-extracted summary. Written by AI from the paper's abstract only. Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.
Improves online multiclass classification with surrogate losses by exploiting prediction stability of margin-based classifiers after parameter updates.
Problem
Online multiclass classification where learners are evaluated through surrogate losses, with the goal of improving on existing regret/mistake bounds.
Approach
Identifies a structural property of margin-based classifiers: on many rounds the predicted label does not change after a parameter update on the current example (stability), and exploits this to get improved surrogate regret bounds and algorithms.
Closest papers by meaning
- A Universal Growth Rate for Learning with Smooth Surrogate Losses (NeurIPS 2024 · similarity 0.90)
- Trading off Consistency and Dimensionality of Convex Surrogates for Multiclass Classification (NeurIPS 2024 · similarity 0.89)
- Bandit and Delayed Feedback in Online Structured Prediction (NeurIPS 2025 · similarity 0.87)
- Online Classification with Predictions (NeurIPS 2024 · similarity 0.86)
- Realizable $H$-Consistent and Bayes-Consistent Loss Functions for Learning to Defer (NeurIPS 2024 · similarity 0.86)
- A General Online Algorithm for Optimizing Complex Performance Metrics (ICML 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.