| opportunity |
location |
|
| 13.40.01.C0291 |
Kirtland Air Force Base, NM 871175776 |
The research topic is centered on development of AI engines (primarily reinforcement learning approaches) to compete in a multi-player, competitive, zero-sum game involving space dynamics and space environment constraints, e.g. power, thermal, communications, etc. The project will involve developing the game itself, using openly available game engines and physics simulations, the development of AI engines using openly available environments, training/playing of games to characterize and analyze AI performance, and documenting results. The work will be collaborative with partners in academia, industry, and within the DoD Operational community.
Looking for candidates with a strong background in AI/ML, particularly in reinforcement learning. Experience in applying to games of any kind or AI/ML interfaces with humans is a plus but not required. Strong software skills are required, particularly Python & Matlab. Experience with modern software development environments & tools is a plus but not required. Position requires US Citizenship - dual Citizens are not eligible. Position will require background check and submission for DoD security clearance (required for computer network access).
References
1. Silver, D., et. al. Mastering the Game of Go with Deep Neural Networks and Tree Searchm Nature, v. 529, pp. 484 -- 489, 2016.
2. https://www.engadget.com/2019/01/24/deepmind-ai-starcraft-ii-demonstration-tlo-mana
3. https://magazine.uc.edu/editors_picks/recent_features/alpha.html
4. N. A. Barriga, M. Stanescu, and M. Buro. Combining strategic learning and tactical search in real-time strategy games. CoRR, abs/1709.03480, 2017.
5. M. Buro, S. Ontan, and M. Preuss. Guest editorial real-time strategy games. IEEE Transactions on Computational Intelligence and AI in Games, 8(4):317--318, Dec 2016.
Space; Autonomy; Decision; Artificial Intelligence; Machine Learning; Game; Game Theory; Reinforcement Learning; Real-time strategy;