Beyond $log^2(T)$ regret for decentralized bandits in matching markets

Authors: Soumya Basu, Karthik Abinav Sankararaman, Abishek Sankararaman

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

Reproducibility Variable Result LLM Response
Research Type Experimental We further demonstrate superiority of our algorithm over existing works through simulations. ... Extensive simulations show, despite the restrictive feedback, phased ETC outperforms CA-UCB in general instances, while UCB-D4 does so under uniqueness consistency.
Researcher Affiliation Industry 1Google, Mountain View, CA, USA 2Facebook, Menlo Park, CA, USA 3Amazon, Palo Alto, CA, USA.
Pseudocode Yes Algorithm 1 Phased ETC Algorithm. ... Algorithm 2 UCB-D4 algorithm (for an agent j)
Open Source Code No The paper does not include any explicit statement or link indicating that the source code for the described methodology is publicly available.
Open Datasets No The paper states: 'We generate random instances to compare the performance...'. This indicates the use of synthetically generated data rather than a publicly available dataset with concrete access information.
Dataset Splits No The paper mentions 'We simulate all the algorithms on the same sample paths, for a total 50 sample paths' but does not specify any training, validation, or test dataset splits.
Hardware Specification No The paper does not provide any specific hardware details such as GPU models, CPU types, or memory used for running the experiments or simulations.
Software Dependencies No The paper mentions baselines (UCB-C, CA-UCB) but does not specify any software names with version numbers or other ancillary software dependencies used in the experiments.
Experiment Setup Yes For UCB-D4, we use β = 1/2K and γ = 2.