Tomorrow's seminar will be BYOL: Bring Your Own Lunch. We haven't gotten our catering set up for the new quarter. Hopefully we will have some small snacks/salad/drinks provided. -Philip On Sun, Jan 6, 2019 at 4:11 PM Philip Garrison <philipmg@cs.washington.edu> wrote:

Please join us Tuesday at noon in *CSE 203* for the first Change talk of the quarter with Dominic Widdows of Grab.

*Who:* Dominic Widdows *What:* Southeast Asia, NLP, and Grab *When:* Tuesday, Jan 8, 12-1pm *Where:* CSE 203

*Abstract* Grab has become the leading transportation company, and increasingly technology platform, throughout Southeast Asia. We have unique challenges and opportunities with language understanding and interfaces. Grab needs to support a bigger range of languages than any of the comparable ridehailing companies, and Southeast Asian languages are much less well-resourced than other world languages that have far fewer speakers. At the same time, we have large amounts of text through reviews, customer service, and chat, and some of the unsupervised learning techniques currently at the forefront of Natural Language Processing and Artificial Intelligence are well-placed to make use of these texts to enhance our language understanding capabilities. Long-term dreams include a seamless way to interact with Grab’s products as if you’re talking to a very knowledgeable concierge. Interim steps include understanding customer service inquiries, translation, and various kinds of sentiment analysis. This talk will be a very brief introduction to the language landscape, NLP techniques, and how we can use them.

*About the speaker* From a background in differential geometry, Dominic started work in NLP in 2001, developing vector space models for language representation and reasoning back when this was still an esoteric and obscure avenue of research. He’s best known in this field for the book Geometry and Meaning, the use of quantum logic to represent search queries and word meanings, and the SemanticVectors package, which has been building word embedding models since 2007 and is freely available on GitHub for anyone to play with. He’s worked on NLP, machine learning, information extraction, and transportation logistics at Google and Microsoft, before joining Grab in 2016, hoping to contribute in some of the world’s most exciting and challenging places _______________________________________________ change mailing list change@change.washington.edu https://changemm.cs.washington.edu/mailman/listinfo/change