The Joint 15th Linguistic Annotation Workshop (LAW) and 3rd Designing Meaning Representations (DMR) Workshop

Co-located with EMNLP 2021 – November 11, 2021

• Home
• Call for Papers
• Program
• Invited talks
• Committees

Invited Speakers

Stephan Oepen

Linguistic Structure beyond Strings and Trees

Abstract: Since 2014, a series of shared tasks at the SemEval, CoNLL, and IWPT conferences has contributed to the advancement of parsing technologies that map natural language utterances to representations of syntactico-semantic structure in the form of general graphs. Efficient and accurate parsers are available for a broad range of linguistic frameworks and a handful of different languages. There has been tremendous progress in methods and techniques, even if the downstream utility of these representations, to some extent, has yet to be demonstrated. I will review some high-level linguistic traits across different frameworks and the evolution of corresponding parsing technologies. I will further relate these developments to recent and ongoing work on capturing structure beyond sentence meaning in general graphs, e.g. negation or structured sentiment analysis.

Bio: Stephan Oepen is professor in language technologies in the Department of Informatics at the University of Oslo. He has studied linguistics, German and Russian philology, computer science, and computational linguistics at Berlin, Volgograd, and Saarbrücken. Prior to his current appointment, he worked, among others, at the German Research Center for Artificial Intelligence and Saarland University (both Germany), YY Technologies (Mountain View), and Stanford University (both USA). His research revolves around syntactico-semantic representations, parsing and generation, grammar engineering, and grammar-based treebanking. https://www.mn.uio.no/ifi/english/people/aca/oe/

Joyce Chai, University of Michigan

Intuitive Physics in Commonsense Language Understanding

Abstract: From a young age, humans acquire commonsense knowledge about a wide variety of physical phenomena, such as movement, rigidity, causes and effects of physical actions. This knowledge, together with our reasoning skills, allows us to understand language and follow instructions effortlessly. However, what seems obvious to humans turns out remarkably difficult for artificial agents. In this talk, I will introduce several efforts in my lab that address understanding and reasoning of intuitive physics in language understanding. I will talk about modeling physical causality in verb semantics for language grounding. I will also introduce a new dataset - Tiered Reasoning for Intuitive Physics (TRIP) - with dense annotations to enable verifiable commonsense reasoning and discuss challenges and opportunities.

Bio: Joyce Chai is a Professor in the Department of Electrical Engineering and Computer Science at the University of Michigan. Prior to joining UM in 2019, she was a Professor of Computer Science at Michigan State University. Her research interests include natural language processing, situated dialogue, human-robot communication, and artificial intelligence. Her recent work explores the intersection of language, vision, and robotics, particularly focusing on grounded language processing to facilitate situated communication with robots and other artificial agents. She has served on the executive board of North America Chapter of Association for Computational Linguistics (NAACL) and as Program Co-chairs for multiple conferences - most recently the 2020 Annual Meeting of Association for Computational Linguistics (ACL). She is a recipient of the National Science Foundation Career Award in 2004, the Best Long Paper Award from ACL 2010, and the William Beal Distinguished Scholar Award from MSU in 2018. She holds a Ph.D. in Computer Science from Duke University.