Deep Reinforcement Learning of Abstract Reasoning from Demonstrations

• Published in: 2018 13th ACM/IEEE International Conference on Human-Robot Interaction (HRI) pp. 160 - 168
• Main Authors: Clark-Turner, Madison, Begum, Momotaz
• Format: Conference Proceeding
• Language: English
• Published: New York, NY, USA ACM 26.02.2018
• Series: ACM Conferences
• Subjects: Abstract Reasoning; Analytical models; Cognition; Computing methodologies > Artificial intelligence > Computer vision > Computer vision tasks > Vision for robotics; Computing methodologies > Artificial intelligence > Knowledge representation and reasoning; Computing methodologies > Artificial intelligence > Knowledge representation and reasoning > Cognitive robotics; Computing methodologies > Machine learning > Learning paradigms > Supervised learning; Deep Learning; Human-robot interaction; Learning from Demonstration; Object recognition; Probabilistic logic; Reinforcement learning; Theory of computation > Logic > Abstraction; Theory of computation > Logic > Automated reasoning; Training; deep learning; abstract reasoning; learning from demonstration
• ISBN: 9781450349536; 1450349536
• ISSN: 2167-2148
• DOI: 10.1145/3171221.3171289

Extracting a set of generalizable rules that govern the dynamics of complex, high-level interactions between humans based only on observations is a high-level cognitive ability. Mastery of this skill marks a significant milestone in the human developmental process. A key challenge in designing such an ability in autonomous robots is discovering the relationships among discriminatory features. Identifying features in natural scenes that are representative of a particular event or interaction (i.e. »discriminatory features») and then discovering the relationships (e.g., temporal/spatial/spatio-temporal/causal) among those features in the form of generalized rules are non-trivial problems. They often appear as a »chicken-and-egg» dilemma. This paper proposes an end-to-end learning framework to tackle these two problems in the context of learning generalized, high-level rules of human interactions from structured demonstrations. We employed our proposed deep reinforcement learning framework to learn a set of rules that govern a behavioral intervention session between two agents based on observations of several instances of the session. We also tested the accuracy of our framework with human subjects in diverse situations.