via MS Teams: Department of Computing and Software, “Candidate Lectures and Seminars” Channel
Knowledge Driven Natural Language Processing
Knowledge is indispensable for humans to understand and produce languages. Likewise natural language processing benefits tremendously from both structured and unstructured knowledge. To this end, my research evolves around an iterative scheme: extract knowledge from large amount of text, and use that knowledge to help understand and generate more text. In this talk, after a brief highlight of my career, I will focus on my selected work on information extraction, text generation and dialogue systems under the above scheme, as well as my recent work on making NLP models smaller and more accessible through model compression. Finally, I will present some ideas of my future work and possible collaborations on light-weight and interdisciplinary NLP.
Kenny Q. Zhu is a full professor of computer science at Shanghai Jiao Tong University (SJTU). He graduated from National University of Singapore with BEng and PhD. Before joining SJTU in 2010, he held positions at Microsoft Redmond and Princeton University. From 2019 to 2021, he was the deputy head of the Computer Science & Engineering Department of SJTU. Kenny’s main research interests are natural language processing and knowledge engineering. He has published extensively in AI, NLP and Database, and serve regularly as SPC or PC at top AI and NLP conferences. His research has been supported by NSF China, MOE China, Microsoft, Google, Oracle, Morgan Stanley, AstraZeneca, Meituan, China Merchant Bank, etc. Kenny is the winner of the 2013 Google Faculty Research Award and 2014 DASFAA Best Paper Award.