Authors:
Daphne Liu
and
Lenhart Schubert
Affiliation:
University of Rochester, United States
Keyword(s):
Self-aware agent, Self-motivated cognitive agent, Introspection.
Related
Ontology
Subjects/Areas/Topics:
Agent Models and Architectures
;
Agents
;
Artificial Intelligence
;
Autonomous Systems
;
Cognitive Systems
;
Computational Intelligence
;
Conversational Agents
;
Evolutionary Computing
;
Formal Methods
;
Informatics in Control, Automation and Robotics
;
Intelligent Control Systems and Optimization
;
Knowledge Representation and Reasoning
;
Planning and Scheduling
;
Simulation and Modeling
;
Soft Computing
;
Symbolic Systems
;
Task Planning and Execution
Abstract:
We present our preliminary work on our framework for building self-motivated, self-aware agents that plan continuously so as to maximize long-term rewards. While such agents employ reasoned exploration of feasible sequences of actions and corresponding states, they also behave opportunistically and recover from failure, thanks to their quest for rewards and their continual plan updates. The framework allows for both specific and general (quantified) knowledge, epistemic predicates such as knowing-that and knowing-whether, for incomplete knowledge of the world, for quantitative change, for exogenous events, and for dialogue actions. Question answering and experimental runs are shown for a particular agent ME in a simple world of roads, various objects, and another agent, demonstrating the value of continual, deliberate, reward-driven planning.