Mitigating Disparity while Maximizing Reward: Tight Anytime Guarantee for Improving Bandits

Authors: Vishakha Patil, Vineet Nair, Ganesh Ghalme, Arindam Khan

IJCAI 2023 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Experimental In this paper, we focused on a self-contained theoretical study of the IMAB problem. Our algorithm provides O(k) compet-itive ratio, whereas HKR can only achieve Θ(k2) competitive ratio (see Appendix G). ...Finally, we further highlight the significance our theoretical results via experiments. ... We compared the performance of the two algorithms on the instances in [Heidari et al., 2016] (see Figure 10). We also compared the performance of our algorithm, HKR, and Round Robin on some randomly generated IMAB instances (see Figure 11).
Researcher Affiliation Collaboration Vishakha Patil1 , Vineet Nair2 , Ganesh Ghalme3 and Arindam Khan1 1Indian Institute of Science, Bangalore 2Arithmic Labs 3Indian Institute of Technology, Hyderabad
Pseudocode Yes Algorithm 1: Horizon-Unaware Improving Bandits ALG
Open Source Code No The paper does not contain an explicit statement or link indicating that the source code for their methodology is publicly available.
Open Datasets No The paper mentions comparing performance 'on the instances in [Heidari et al., 2016]' and 'randomly generated IMAB instances', but does not provide concrete access information (e.g., specific links, DOIs, or formal citations to public datasets) for a publicly available dataset used for training.
Dataset Splits No The paper does not provide specific details about training/validation/test dataset splits. The problem context (IMAB) involves time steps and pulls, not traditional dataset splits.
Hardware Specification No The paper does not specify any hardware used for running the experiments (e.g., GPU/CPU models, memory, or cloud instances).
Software Dependencies No The paper does not list specific software dependencies with version numbers.
Experiment Setup No The paper describes the algorithms and their theoretical properties but does not provide specific experimental setup details such as hyperparameter values, learning rates, or specific training configurations for the experiments mentioned in Section 5 and Appendix F.