Cognitive Architecture for Introspecting Deep Reinforcement Learning Agents
Loading...
Authors
Mitsopoulos, Konstantinos
Somers, Sterling
Lebiere, Christian
Thomson, Robert
Issue Date
2020
Type
Conference presentations, papers, posters
Language
Keywords
deep reinforcement learning , machine learning , cognitive model , cognitive architecture , explainable artificial intelligence
Alternative Title
Abstract
In this work we demonstrate the use of Cognitive Architectures in the Reinforcement Learning domain, specifically to serve as a common ground to understand and explain Reinforcement Learning agents in Human Ontology terms. Cognitive Architectures could potentially act as an adaptive bridge between Cognition and modern AI, sensitive to the cognitive dynamics of human user and the learning dynamics of AI agents.
Description
Citation
Mitsopoulos, Konstantinos, Sterling Somers, R. Thomson, and C. Lebiere. "Cognitive architectures for introspecting deep reinforcement learning agents." In Workshop on Bridging AI and Cognitive Science, at the 8th International Conference on Learning Representations. ICLR. 2020.
Publisher
ICLR 2020