Intelligent information systems often harvest world knowledge in the form of information tuples, which are stored in knowledge-bases (KBs). These KBs power downstream information tasks. In this talk I will briefly describe various threads of research that interact with KBs in different ways. These include (1) extracting information to populate a KB, (2) inferring unstated information to improve the completeness of KBs and (3) using these KBs for downstream tasks such as goal-oriented dialog. In the two part talk, I will first summarize about a decade of progress on the Open Information Extraction project, and later discuss some of the more recent research on neural models for inference and dialog.
Prof. Mausam is an associate professor at the Dept. of Computer Science in IIT-Delhi and an affiliate faculty at the University of Washington. He got his undergraduate degree from IIT-Delhi in Computer Science in 2001 and completed his PhD. work titled Stochastic Planning with Concurrent, Durative Actions from the University of Washington in 2007.
Mausam’s work focuses on large-scale information extraction and text summarization, AI & ML applications to crowdsourcing and education, automated planning under uncertainty, machine learning, and probabilistic reasoning.