ART-AI Seminar
We are pleased to have Mohan Sridharan, a Reader in Cognitive Robot Systems in the School of Computer Science at the University of Birmingham join us for this ART-AI seminar, entitled ‘Towards Integrated and Transparent Reasoning, Control, and Learning in Cognitive Robotics’
The seminar will take place online on Zoom on Tuesday 16th November 13.15pm-14.15pm and will be chaired by Marina De Vos.
Abstract
In this talk, I will describe ongoing work on a refinement-based architecture for transparent reasoning, learning, and control in robotics. The architecture combines the complementary strengths of knowledge-based reasoning, and data-driven learning and control. Specifically, the architecture represents and reasons with non-monotonic logic-based and probabilistic descriptions of incomplete common-sense domain knowledge at different tightly-coupled abstractions. Reasoning triggers and guides cumulative learning of previously unknown domain knowledge (when needed) based on deep learning, reinforcement learning, and inductive learning methods. Furthermore, the interplay between representation, reasoning, learning, and control is used to enable the robot to provide on-demand relational descriptions of its decisions, beliefs, and experiences. The capabilities of the architecture are demonstrated in the context of simulated and physical robots assisting humans in complex domains.
Bio
Mohan Sridharan is a Reader in Cognitive Robot Systems in the School of Computer Science at the University of Birmingham (UK). He was previously a Senior Lecturer in the Department of Electrical and Computer Engineering at The University of Auckland (NZ) and prior to that, he was a faculty member at Texas Tech University. Mohan Sridharan received his Ph.D. in Electrical and Computer Engineering from The University of Texas at Austin. His primary research interests include knowledge representation and reasoning, machine learning, computer vision and cognitive systems as applied to autonomous robots and adaptive agents. He develops architectures and algorithms that enable robots to collaborate with non-expert human participants, acquiring and using sensor inputs and high-level human feedback based on need and availability. Furthermore, he is interested in developing algorithms for estimation and prediction problems in application domains such as agricultural irrigation management, intelligent transportation, and climate informatics.
