Convex-Concave Zero-Sum Markov Stackelberg Games

Authors: Denizalp Goktas, Arjun Prakash, Amy Greenwald

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

Reproducibility Variable Result LLM Response
Research Type Experimental We also prove that reach-avoid problems are naturally modeled as convex-concave zero-sum Markov Stackelberg games, and show experimentally that Stackelberg equilibrium policies are more effective than their Nash counterparts in these problems.1
Researcher Affiliation Academia Denizalp Goktas Brown University, Computer Science denizalp_goktas@brown.edu Arjun Prakash Brown University, Computer Science arjun_prakash@brown.edu Amy Greenwald Brown University, Computer Science amy_greenwald@brown.edu
Pseudocode Yes Algorithm 1 Saddle-Point-Oracle SGD/Nested SGDA
Open Source Code Yes Our code is found at: https://github.com/arjun-prakash/stackelberg-reach-avoid.
Open Datasets No The paper describes the setup for a reach-avoid game with specific parameters but does not refer to a publicly available dataset by name or provide access information (link, DOI, or specific citation).
Dataset Splits No The paper describes running '100 games' for evaluation but does not specify dataset splits (e.g., percentages or counts for training, validation, or testing).
Hardware Specification No The paper does not provide specific hardware details such as GPU/CPU models, processor types, or memory used for running its experiments.
Software Dependencies No The paper does not provide specific software dependencies with version numbers (e.g., library or solver names with versions).
Experiment Setup Yes Our experiments were run on a 7x7 square grid, with the target set T a closed ball of radius 1 centered along the lower edge, and the avoid set V a closed ball of radius 0.3 around the antagonist. We set the bonus (resp. penalty) for reaching the target (resp. avoid set) β = 200, ! = 30 , and = 0.25.