Research Projects

Human reasoning has been a subject of interest for centuries, with studies focusing on visual behavior during problem-solving. However, there is a lack of literature on how humans tackle cognitive problems and how AI engines solve them. Learning is a common form of problem-solving, with human information processing being influenced by limited-capacity short-term memory systems and unlimited long-term memory systems. There is a conceptual gap between human problem-solving and machine models, often attributed to human information processing and long-short-term memory features. To understand and improve human problem-solving abilities, it is crucial to develop an equivalent computational model that can aid the problem-solving process. This project aims to discover and analyze deeper patterns in human problem-solving processes based on features obtained from humans solving cognitive problems in different platforms. A cognitive-AI-based learning analytics and assistive tool will be developed to visualize improvements in human problem-solving processes. This tool will provide a platform to incorporate AI-driven methods to understand problem-solving behaviors in visuo-spatial problems with the assistance of multimodal sensors. The tool framework will surface automated cues as feedback to the human problem-solver, guiding the repair and improvement of the conceptual learning process. Eye-tracking movements and suitable metrics will be used to draw conclusions about the improvement in the learning process.