Research map / Bandits and online learning / Contextual bandits and Thompson sampling
Provably Efficient Multi-Task Meta Bandit Learning via Shared Representations
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.
Studies multi-task meta bandit learning with shared low-dimensional representations, providing theory on how representation learning improves sample efficiency for new bandit tasks.
Problem
Meta-learning for bandits: how representation learning across multiple tasks can transfer knowledge to improve learning of new tasks in data-scarce settings.
Approach
Uses shared representations across bandit tasks to enable provably efficient multi-task meta learning; the abstract is truncated so the specific algorithm is not stated.
Closest papers by meaning
- Meta-learning with Stochastic Linear Bandits (ICML 2020 · similarity 0.92)
- Impact of Representation Learning in Linear Bandits (ICLR 2021 · similarity 0.92)
- Fast and Sample Efficient Multi-Task Representation Learning in Stochastic Contextual Bandits (ICML 2024 · similarity 0.90)
- Beyond task diversity: provable representation transfer for sequential multitask linear bandits (NeurIPS 2024 · similarity 0.90)
- A Classification View on Meta Learning Bandits (ICML 2025 · similarity 0.89)
- Reciprocal Learning (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.