No-Regret Learning for Fair Multi-Agent Social Welfare Optimization

Authors: Mengxiao Zhang, Ramiro Deo-Campo Vuong, Haipeng Luo

NeurIPS 2024 | Conference PDF | Archive PDF | Plain Text | LLM Run Details

Reproducibility Variable Result LLM Response
Research Type Theoretical In this work, we provide a complete answer to this question in various settings. Specifically, in stochastic N-agent K-armed bandits, we develop an algorithm with e O(K 2 N T N 1 N ) regret and prove that the dependence on T is tight... we prove that no algorithm can achieve sublinear regret. To circumvent such negative results, we further consider a setting with full-information feedback and design two algorithms with T-regret
Researcher Affiliation Academia Mengxiao Zhang University of Iowa mengxiao-zhang@uiowa.edu Ramiro Deo-Campo Vuong Cornell University ramdcv@cs.cornell.edu Haipeng Luo University of Southern California haipengl@usc.edu
Pseudocode Yes Algorithm 1 UCB for N-agent K-armed NSW maximization... Algorithm 2 FTRL for N-agent K-armed SWF maximization with full-info feedback
Open Source Code No No explicit statement about releasing code or links to a repository was found. The NeurIPS checklist mentions: 'This paper does not release new assets.'
Open Datasets No This paper does not include experiments. Therefore, no dataset information for training, validation, or testing is provided.
Dataset Splits No This paper does not include experiments. Therefore, no dataset information for training, validation, or testing is provided.
Hardware Specification No This paper does not include experiments, and therefore no hardware specifications are provided.
Software Dependencies No No specific ancillary software details with version numbers were mentioned in the paper.
Experiment Setup No This paper does not include experiments or specific experimental setup details like hyperparameters or training configurations.