PulseAugur
EN
LIVE 19:29:22

New research details sample complexity for bandit learning algorithms

By PulseAugur Editorial · [5 sources] ·

Two new research papers explore the sample complexity of bandit learning algorithms in multiclass classification settings. The first paper introduces the "bandit DS dimension" to characterize sample complexity for PAC learning with bandit feedback, proposing a new algorithm called ListCascade. The second paper focuses on sparse contextual bandits, presenting algorithms that achieve improved sample complexity bounds by leveraging the sparsity of reward vectors and analyzing the decision-estimation coefficient (DEC). Both papers aim to provide tighter theoretical guarantees for learning in scenarios where full label information is not available. AI

IMPACT These papers offer theoretical advancements in bandit learning, potentially leading to more efficient algorithms for tasks with limited feedback.

RANK_REASON The cluster contains two academic papers published on arXiv detailing theoretical advancements in machine learning algorithms.

Read on arXiv cs.LG →

AI-generated summary · Google Gemini · from 5 sources. How we write summaries →

New research details sample complexity for bandit learning algorithms

COVERAGE [5]

  1. arXiv cs.LG TIER_1 English(EN) · Steve Hanneke, Qinglin Meng, Shay Moran, Amirreza Shaeiri ·

    PAC Learning with Bandit Feedback: Sharp Sample Complexity in the Realizable Setting

    arXiv:2605.25678v1 Announce Type: cross Abstract: We study the problem of multiclass PAC learning with bandit feedback in the realizable setting. In this framework, there is an unknown data distribution over an instance space $\mathcal{X}$ and a label space $\mathcal{Y}$, as in c…

  2. arXiv stat.ML TIER_1 English(EN) · Liad Erez, Fan Chen, Alon Cohen, Tomer Koren, Yishay Mansour, Shay Moran, Alexander Rakhlin ·

    The Sample Complexity of Multiclass and Sparse Contextual Bandits

    arXiv:2605.29645v1 Announce Type: cross Abstract: We study contextual bandits in the stochastic i.i.d.\ setting, where a learner observes contexts drawn from an unknown distribution, selects actions from a finite set $A$, and aims to identify an approximately optimal policy from …

  3. arXiv stat.ML TIER_1 English(EN) · Alexander Rakhlin ·

    The Sample Complexity of Multiclass and Sparse Contextual Bandits

    We study contextual bandits in the stochastic i.i.d.\ setting, where a learner observes contexts drawn from an unknown distribution, selects actions from a finite set $A$, and aims to identify an approximately optimal policy from a given class based on bandit feedback. Motivated …

  4. arXiv stat.ML TIER_1 English(EN) · Amirreza Shaeiri ·

    PAC Learning with Bandit Feedback: Sharp Sample Complexity in the Realizable Setting

    We study the problem of multiclass PAC learning with bandit feedback in the realizable setting. In this framework, there is an unknown data distribution over an instance space $\mathcal{X}$ and a label space $\mathcal{Y}$, as in classical multiclass PAC learning, but the learner …

  5. arXiv stat.ML TIER_1 English(EN) · Amirreza Shaeiri ·

    PAC Learning with Bandit Feedback: Sharp Sample Complexity in the Realizable Setting

    We study the problem of multiclass PAC learning with bandit feedback in the realizable setting. In this framework, there is an unknown data distribution over an instance space $\mathcal{X}$ and a label space $\mathcal{Y}$, as in classical multiclass PAC learning, but the learner …

RELATED ENTITIES

RELATED TOPICS