We are excited to host John Langford speaking on Real World Reinforcement Learning
We’ve created a dozen real world reinforcement learning deployments enabling use cases around contextual personalization and contextual optimization culminating in the new Cognitive Services Personalizer (https://azure.microsoft.com/en-us/services/cognitive-services/personalizer/) which makes these techniques available to the world on a mass scale. I’ll discuss the key technology (and papers) behind this along with the possibilities that it allows. Amongst other things, the counterfactual evaluation system we’ve created makes optimization over policies radically faster and friendlier than A/B testing.
John Langford is a machine learning research scientist, a field which he says "is shifting from an academic discipline to an industrial tool". He is the author of the weblog http://hunch.net/ and the principal developer of Vowpal Wabbit (http://hunch.net/~vw/). John works at Microsoft Research New York, of which he was one of the founding members.
**Doors open at 6:30 pm**; the presentations will begin right around 7:00 pm; and, yes, there will be refreshments of all kinds and pizza.
You'll have to check-in with security with your Name/ID. Definitely sign-up if you’re going to attend–unfortunately people whose names aren’t entered into the security system in advance won’t be allowed in.
After John's presentation, we will open up the floor to discussion and questions.
**Talks are always recorded on video and released ~2 weeks after the meetup.**
We hope that you'll read some of the papers and references before the meetup, but don't stress if you can't. If you have any questions, thoughts, or related information, please visit #pwlnyc (https://paperswelove.slack.com/messages/pwlnyc/) on slack (http://papersweloveslack.herokuapp.com/), our GitHub repository (https://github.com/papers-we-love/papers-we-love), or add to the discussion on this event's thread.