A Platform for Engineering Meta-Cognitively Perceptive Artificial Teammates
Abstract
Most real and robust military missions begin with a pristine plan, but soon the environment has its sobering say, and success or fai,lure then hinges on the level of adaptability the agents in question can muster on the fly. Therefore, artificial agents, whether p,hysical robots or purely digital ones in cyberspace, will not be maximally effective teammates with human soldiers unless these agen,ts are capable of sustained, dynamic, defeasible (i.e. nonmonotonic), provably correct meta-reasoning in these in-flux missions. We, refer to this type of reasoning, which has yet to be formalized, let alone implemented, as *M*-reasoning. In particular, when the,mission environment severely compromises the perception of artificial teammates, this powerful new brand of meta-reasoning is demand,ed. By further advancing, and deeply integrating, two established, best-in-breed frameworks each of our respective groups has previ,ously erected --- a reasoning-centric ensemble of unprecedentedly expressive cognitive calculi and automated-reasoning technology fr,om the RAIR-Lab; and a perception-centric computational architecture of human cognition (ARCADIA) from the NRL side --- we will demo,nstrate and deliver a novel platform for the engineering of artificial teammates capable of *M*-reasoning.Approved for Public Releas,e
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- May 16, 2022
- Source ID
- N000142212201
Entities
People
- Selmer Bringsjord
Organizations
- Office of Naval Research
- Rensselaer Polytechnic Institute
- United States Navy