Commonsense and World Knowledge

Everyday intelligence relies on a large store of background—or “common sense”—knowledge about the world. As humans, we acquire this knowledge in part through our experiences; artificially intelligent systems currently do not have access to this same kind of input. However, both explicit and implicit information about the world can be learned from data available in text and images. This session focuses on some key research in this area, with talks on extracting information from data in order to acquire commonsense knowledge about the human world.

Speaker Details

“Margaret Mitchell is a researcher in Microsoft’s NLP Research Group, working on grounded language generation. Before joining Microsoft, she was a postdoctoral researcher at The Johns Hopkins University Center of Excellence, where she worked on semantic role labeling and sentiment analysis using graphical models. ”

Raymond J. Mooney is a Professor in the Department of Computer Science at the University of Texas at Austin. He received his Ph.D. in 1988 from the University of Illinois at Urbana/Champaign. He is an author of over 150 published research papers, primarily in the areas of machine learning and natural language processing. He was the President of the International Machine Learning Society from 2008–2011, program co-chair for AAAI 2006, general chair for HLT-EMNLP 2005, and co-chair for ICML 1990. He is a Fellow of the American Association for Artificial Intelligence, the Association for Computing Machinery, and the Association for Computational Linguistics, and the recipient of best paper awards from AAAI-96, KDD-04, ICML-05 and ACL-07.

Dan Roth is a Professor in the Department of Computer Science and the Beckman Institute at the University of Illinois at Urbana-Champaign and a University of Illinois Scholar. He is the director of the DHS funded Center for Multimodal Information Access & Synthesis (MIAS) and has faculty positions also at the Statistics, Linguistics and ECE Departments and at the graduate School of Library and Information Science.

Lucy Vanderwende’s research focuses on text understanding. She is deeply involved with developing MindNet, a method for automatically acquiring semantic information; when extracting from dictionaries, much of this information might be considered common sense. Lucy has been involved in projects to extract information from electronic health records, for phenotype determination and prediction, where identifying the truth value of the sentence contributes to greater accuracy. Further, Lucy has worked with Sumit Basu on various projects of Question Generation, primarily for the purpose of generating quizzes from arbitrary text to support self-motivated learning. Lucy has been at Microsoft Research since 1992 and Affiliate Faculty at University of Washington Department of Biomedical Health Informatics since 2011.

Benjamin Van Durme is the lead of Natural Language Understanding research at the Johns Hopkins University Human Language Technology Center of Excellence (HLTCOE), as well as an assistant research professor in both Computer Science and Cognitive Science at Johns Hopkins University. Between the HLTCOE, his small army of PhD students, and affiliated post-doctoral researchers, research areas covered include: information extraction and knowledge base population, streaming and randomized algorithms in HLT, topic modeling, author attribute prediction in social media, educational NLP, and various aspects of computational semantics.

Series:: Microsoft Research Faculty Summit
Date:: July 24, 2015
Speakers:: Benjamin Van Durme, Dan Roth, Lucy Vanderwende, Margaret Mitchell, and Raymond J. Mooney
Affiliation:: Microsoft Research, University of Texas at Austin, University of Illinois at Urbana-Champaign, Johns Hopkins University