We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

cs.LG

Change to browse by:

References & Citations

DBLP - CS Bibliography

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo ScienceWISE logo

Computer Science > Machine Learning

Title: Inverse Reinforcement Learning for Strategy Identification

Abstract: In adversarial environments, one side could gain an advantage by identifying the opponent's strategy. For example, in combat games, if an opponents strategy is identified as overly aggressive, one could lay a trap that exploits the opponent's aggressive nature. However, an opponent's strategy is not always apparent and may need to be estimated from observations of their actions. This paper proposes to use inverse reinforcement learning (IRL) to identify strategies in adversarial environments. Specifically, the contributions of this work are 1) the demonstration of this concept on gaming combat data generated from three pre-defined strategies and 2) the framework for using IRL to achieve strategy identification. The numerical experiments demonstrate that the recovered rewards can be identified using a variety of techniques. In this paper, the recovered reward are visually displayed, clustered using unsupervised learning, and classified using a supervised learner.
Comments: The paper has been accepted as a regular paper in IEEE International Conference on Systems, Man, and Cybernetics (SMC), 2021
Subjects: Machine Learning (cs.LG); Artificial Intelligence (cs.AI); Computer Science and Game Theory (cs.GT)
Cite as: arXiv:2108.00293 [cs.LG]
  (or arXiv:2108.00293v1 [cs.LG] for this version)

Submission history

From: Mark Rucker [view email]
[v1] Sat, 31 Jul 2021 17:22:52 GMT (1253kb,D)

Link back to: arXiv, form interface, contact.