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.


On Sun, Jan 6, 2019 at 4:11 PM Philip Garrison <>

> 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 mailing list

Reply via email to