Dr. Unmesh Kurup
Unmesh Kurup, Ph.D.
is Post-doctoral Researcher, Department of Cognitive Science,
Rensselaer Polytechnic Institute.
His areas of specialization include: multi-modal cognition, cognitive architectures, diagrammatic reasoning, spatial reasoning, and knowledge representation.
Human reasoning uses many non-symbolic representations, the most common one being some sort of diagram or sketch. There is consensus that such representations (especially such external representations) play an important role in the reasoning process though their exact representational nature is a matter of contention. The debates between Pylyshyn and Kossyln notwithstanding, Unmesh’s approach has been to study these problems from a computational perspective and within the constraints of an architectural framework. Integration lies at the core of this process because it involves adding additional modalities while minimizing changes to the existing structure.
Most of Unmesh’s research has been focused on the use of nonsymbolic representations in problem solving, but these representations are useful in many other situations. For example, he has investigated their use in cognitive modeling tasks where their use can be shown to result in errors in recall. In addition, there are also applications to HCI where the ability to represent and use nonsymbolic representations allows an artificial agent to effectively communicate with humans. This and other advantages (use in recall, episodic memory, etc.) of such representations are additional areas of interest.
The overall goal of his research to is to understand the nature of nonsymbolic representations such as diagrammatic or spatial representations and their role in human cognition, especially in problem solving. He’s particularly attracted to the cognitive architecture approach to studying these problems due to a number of reasons including the fact that they provide a baseline against which to compare and contrast the effectiveness of such representations.
He coauthored Quantitative Spatial Reasoning for General Intelligence, Representational and Inferential Requirements for Diagrammatic Reasoning in the Entity Re-Identification Task, and Integrating Perception and Cognition for AGI, A Cognitive Map for an Artificial Agent, Integrating Constraint Satisfaction and Spatial Reasoning, Multi-modal Cognitive Architectures: A Partial Solution to the Frame Problem, Diagrammatic Reasoning in Support of Situation Understanding and Planning, and An Architecture for Adaptive Algorithmic Hybrids.
Unmesh earned his Ph.D. in Computer Science at Ohio State University in 2008 where he specialized in AI, Computer Science, and Engineering. His dissertation was Design and Use of a Bimodal Cognitive Architecture for Diagrammatic Reasoning and Cognitive Modeling.
Watch A Cognitive Map for an Artificial Agent and Unmesh Kurup, Post Doc, RPI.