A virtual reality (VR) environment for studying temporal decision making and signalling between a human and an artifical intelligence (AI) agent, characterized by a sparse reward reinforcement learning problem, audio+visual+vibrotactile stimulus, and partially observable domain mechanics.
Please note: this project requires SteamVR or another VR package set for Unity; specific SteamVR settings and files have been excluded from this repository, so some work may be required to re-insert the VR rigging approach when deploying the project on a new machine or with new hardware. Importantly, the main experimental code is attached to the camera gameObject inside the VR Rig, so please note this will need to be moved to a new rig if one is used. Similarly, updates may be required to bring this project up to date with current Unity LTS version used on the target machine.
These project files were the basis for the VR experiments on Pavlovian signalling during multiagent interaction in the following manuscripts:
- Patrick M. Pilarski, Andrew Butcher, Elnaz Davoodi, Michael Bradley Johanson, Dylan J. A. Brenneis, Adam S. R. Parker, Leslie Acker, Matthew M. Botvinick, Joseph Modayil, Adam White (2022). The Frost Hollow Experiments: Pavlovian Signalling as a Path to Coordination and Communication Between Agents. arXiv:2203.09498 [cs.AI]
- Dylan J. A. Brenneis, Adam S. Parker, Michael Bradley Johanson, Andrew Butcher, Elnaz Davoodi, Leslie Acker, Matthew M. Botvinick, Joseph Modayil, Adam White, Patrick M. Pilarski (2022). Assessing Human Interaction in Virtual Reality With Continually Learning Prediction Agents Based on Reinforcement Learning Algorithms: A Pilot Study. Adaptive and Learning Agents (ALA) Workshop at AAMAS 2022, Auckland, NZ, 9-10 May 2022.
- Andrew Butcher, Michael Bradley Johanson, Elnaz Davoodi, Dylan J. A. Brenneis, Leslie Acker, Adam S. R. Parker, Adam White, Joseph Modayil, Patrick M. Pilarski (2022) Pavlovian signalling with general value functions in agent-agent temporal decision making. Adaptive and Learning Agents (ALA) Workshop at AAMAS 2022, Auckland, NZ, 9-10 May 2022.