Research map / Bandits and online learning / Online convex optimization and games
A Theory of Adversary-Directed Online Learning
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.
Introduces adversary-directed online learning, where the learner knows the instance set in advance but an adversary adaptively chooses the presentation order, and develops a theory for it.
Problem
Online learning where the learner knows the set of instances beforehand but an adversary adaptively determines their ordering; this formulation has not been previously studied, unlike traditional adversarial online learning.
Approach
Formalizes a new online learning model in which the instance set is known to the learner and only the adversary's ordering of instances is adaptive, distinguishing it from standard adversarial online learning and related models such as transductive online learning.
Experimental setup
Builds on: transductive online learning
Closest papers by meaning
- Online Learning in the Random-Order Model (ICML 2025 · similarity 0.90)
- Private Online Learning against an Adaptive Adversary: Realizable and Agnostic Settings (NeurIPS 2025 · similarity 0.89)
- Bandit-Feedback Online Multiclass Classification: Variants and Tradeoffs (NeurIPS 2024 · similarity 0.88)
- Strategic Littlestone Dimension: Improved Bounds on Online Strategic Classification (NeurIPS 2024 · similarity 0.88)
- From Average Sensitivity to Small-Loss Regret Bounds under Random-Order Model (NeurIPS 2026 · similarity 0.88)
- Optimal Algorithms for Online Convex Optimization with Adversarial Constraints (NeurIPS 2024 · similarity 0.88)
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.