Here at Scientific Revenue we enjoy sharing what we learn. We usually do this in the form of interacting with customers, writing articles for industry publications, and speaking at industry events. It is extra special when we are invited to give a colloquium talk at a major university.
On November 11, 2016, our CEO Bill Grosso spoke at the University of San Francisco Seminar Series in Analytics, where he presented “Getting to Continuous Optimization.”
The University of San Francisco (USF) runs a weekly Friday afternoon seminar that is open to the community (attendance includes both graduate students and industry professionals). The talks cover a wide range: from open research topics to the nuts and bolts of daily practice.
In his talk, Bill addressed ways to transition from simple A/B testing to continuous optimization, and how you need to go outside the 'mainstream' of data science to build systems that constantly and dynamically adapt to the end user. He also went through some of the less well known tools from the machine learning toolkit while sharing examples from mobile gaming.
We asked Bill how it went:
"Great. I wanted to expose the audience to some of the core ideas from reinforcement learning, and give them some intuition about situations where the standard methods and ideas don't work. It's easy to find out about Logistic Regression and Random Forests. But almost no-one knows about Boltzmann's equation, or entropy, or causal inference. And, while it's hard to go deep in a single hour, it's exciting to be able to point some really smart students in the right direction. I'm looking forward to them asking more questions by email in the coming weeks."
Slides for Bill’s talk can be found here.