Funded Projects

Wikipedia, Udacity, Coursera, edX, Khan Academy, YouTube and others contain more information about any given topic than we can aspire to learn in a lifetime. Such content has the ability dramatically transform our lifelong ability to learn. However, this utopian dream of lifelong and personalized learning is yet to be realized. We believe that the biggest impediment in realizing this transformation is the lack of structure in the content as well as a lack of our ability to model human learning in both short and long term. This BlueSky proposal aims to sow the seeds of a much larger research effort that will undertake the challenge of personalized life-long learning through three main research thrusts Content Synthesis - Through multimodal information retrieval, knowledge representation and unsupervised learning algorithms, we will structurally organize and categorize learning materials in text, natural language, video and other media. Human Perception and Modeling - Through a diverse set of sensors, we will perceive and analyze the behavior of individual and collaborative learning, especially in topics that require both theoretical and hands-on skill sets. AI-Assisted Learning - We will combine the structured content and the human modeling to adaptively provide learning materials to the individual, measure learning outcomes and evolve the learning experience Since this larger future project is ambitious in its vision, it is critical to generate tangible rationale supporting such a structured intelligent approach to lifelong personalized learning. The objective of this BlueSky proposal will be to explore thrust 1 in the context of the contemporary subject of Robot Autonomy as a course taught in the School of Engineering. With this subject as the case study, we will identify potential gaps in our understanding of the problem challenges and potential solutions to structured content synthesis.