Optimal Gradient-based Algorithms for Non-concave Bandit Optimization

Authors: Baihe Huang, Kaixuan Huang, Sham Kakade, Jason D. Lee, Qi Lei, Runzhe Wang, Jiaqi Yang

NeurIPS 2021 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Theoretical In this paper, we design stochastic zeroth-order gradient-like2 ascent algorithms to attain minimax regret for a large class of structured polynomials. ... If you ran experiments... [N/A]
Researcher Affiliation Collaboration 1Peking University 2Princeton University 3University of Harvard 4Microsoft Research 5Tsinghua University
Pseudocode Yes Algorithm 1 Noisy power method for bandit eigenvalue problem. ... As presented in Algorithm 4, we conduct noisy subspace iteration... More details are deferred to Algorithm 5 and the Appendix. We show the bandit optimization procedure in Algorithm 6
Open Source Code No The paper is theoretical and does not mention releasing source code. The questionnaire in the paper explicitly states 'N/A' for code availability in the 'If you ran experiments' section.
Open Datasets No This is a theoretical paper that defines abstract 'action sets' and 'function classes' but does not utilize or refer to specific publicly available datasets for experimental training.
Dataset Splits No This is a theoretical paper and does not involve experimental validation with dataset splits.
Hardware Specification No The paper is theoretical and does not report on experimental hardware specifications. The 'If you ran experiments...' section of the self-evaluation checklist explicitly marks hardware details as 'N/A'.
Software Dependencies No The paper is theoretical and does not report on software dependencies for experimental replication. The 'If you ran experiments...' section of the self-evaluation checklist explicitly marks software details as 'N/A'.
Experiment Setup No The paper is theoretical and does not provide details about an experimental setup, hyperparameters, or system-level training settings. The 'If you ran experiments...' section of the self-evaluation checklist explicitly marks these details as 'N/A'.