NASA Logo

NTRS

NTRS - NASA Technical Reports Server

Collections About News Help Login
Back to Results
Modeling Humans as Reinforcement Learners: How to Predict Human Behavior in Multi-Stage GamesThis paper introduces a novel framework for modeling interacting humans in a multi-stage game environment by combining concepts from game theory and reinforcement learning. The proposed model has the following desirable characteristics: (1) Bounded rational players, (2) strategic (i.e., players account for one anothers reward functions), and (3) is computationally feasible even on moderately large real-world systems. To do this we extend level-K reasoning to policy space to, for the first time, be able to handle multiple time steps. This allows us to decompose the problem into a series of smaller ones where we can apply standard reinforcement learning algorithms. We investigate these ideas in a cyber-battle scenario over a smart power grid and discuss the relationship between the behavior predicted by our model and what one might expect of real human defenders and attackers.
Document ID
20120004027
Acquisition Source
Ames Research Center
Document Type
Conference Paper
Authors
Lee, Ritchie (Carnegie-Mellon Univ. Moffett Field, CA, United States)
Wolpert, David H. (NASA Ames Research Center Moffett Field, CA, United States)
Backhaus, Scott (Los Alamos National Lab. NM, United States)
Bent, Russell (Los Alamos National Lab. NM, United States)
Bono, James (American Univ. Washington, DC, United States)
Tracey, Brendan (Stanford Univ. Stanford, CA, United States)
Date Acquired
August 25, 2013
Publication Date
December 16, 2011
Subject Category
Statistics And Probability
Report/Patent Number
ARC-E-DAA-TN4356
Meeting Information
Meeting: Twenty-Fifth Annual Conference on Neural Information Processing Systems (NIPS2011)
Location: Granada
Country: Spain
Start Date: December 11, 2011
End Date: December 17, 2011
Sponsors: Neural Information Processing Systems Foundation
Funding Number(s)
CONTRACT_GRANT: NNA08CG83C
Distribution Limits
Public
Copyright
Public Use Permitted.

Related Records

There are no records associated with this record.
No Preview Available