Summaries - Office of Research & Innovation
Back Modeling Large-Scale Warfighter Cognitive Reasoning and Decision Making Using Machine Learning, Artificial Intelligence and Game Theory (Continuation)
|Division||Research & Sponsored Programs|
|Department||NPS Naval Research Program|
|Sponsor||NPS Naval Research Program (Navy)|
This IREF proposes a continuation and expansion of the 2018 research NPS-18-N013-A, "Reinforcement Learning for Modeling Large Scale Cognitive Reasoning.”
We continue innovative models of cognitive reasoning and decision making of warfighters using AI/ML. Cognitive architecture, algorithms, and software systems are important to model complex reasoning, cognitive functions and decision making in warfighting environments. For example, reinforcement learning (RL) is a fascinating and active subfield of machine learning (ML) and artificial intelligence (AI) that focuses on training artificial agents to perform cognitive reasoning and decision making to connect uncertain and unknown situations to actions that maximize life-time reward from the environment. The state-of-art ML/AI/GT has been used in the commercial applications for breakthrough applications such as DeepMind’s AlphaGo and AlphaZero (machine vision and RL based ML/AI), CMU’s libratus (game theory based ML/AI). These AIs demonstrated cognitive reasoning and decision making which can beat human masters.
There are profound implications to apply the ML/AI/GT techniques to military applications to achieve the decision making superiority such as modeling, simulation and ROI calculation in the vast complex and uncertain areas of Cybersecurity and Information Warfare context including applications of combat identification, Battlespace Awareness, C-C4ISR, Assured C2, modeling/simulation and war games. To advance the cognitive modeling for warfighters, we will demonstrate deep analytics, ML/AI systems are able to learn from the warfighters and later become their assistants in a cognitive stressful environment to reduce their cognitive burden.
Deliverables include a research report and presentation; student thesis and a publication to a journal or conference.
|Publications||Publications, theses (not shown) and data repositories will be added to the portal record when information is available in FAIRS and brought back to the portal|