Speaker: Winston Saunders
Abstract: Word vectors, derived by deep learning algorithms applied to billions of words of text, provide powerful semantic models of language. Code in R, demonstrating [queen] + [man] - [woman] ~ [King] to about 90% accuracy will be reviewed. Building first on exploratory "bag of words" analysis of Presidential debate texts, we'll explore, using pre-computed GloVe vectors (Pennington et al http://nlp.stanford.edu/projects/glove/ ), relationships like [sanders] + [trump] - [clinton] ~ [cruz] and how candidate positions align to rhetorical sentiment like [government] + [people] - [tax]. This analysis is work in progress. We'll also test empirical limits (aka failed experiments). Active feedback is both sought and welcome.
Doors open ~6 pm, talk starts at 6:30 pm
Doors are open at bottom, take elevator to 3rd floor, door should be open for suite 320
We'll visit a local watering hole afterwards.
Hashtag for PDX R meetups: #pdxrlang (https://twitter.com/hashtag/pdxrlang?src=hash) & the Twitter account to follow/tweet at is @pdxrlang (https://twitter.com/pdxrlang) We also use http://pdxdata.slack.com/ for a back-channel during MeetUps and in between. Invite yourself here: http://pdxdata.org/slack/ ...