Machine learning that matters

This is a past event

200 people went

Location visible to members

What we'll do

For our last meetup for 2015 we are extremely fortunate to have two leaders in the applied machine learning field to present to us. Hope you can join us for what will be a great night. Again Optiver have been kind enough to sponsor great beer, pizza and venue for the night.

Please only RSVP if you are sure you can come so that someone on the waitlist does not miss out.

Which Learning Algorithms Really Matter (Industrially)? Ted Dunning, Chief Application Architect, MapR Technologies (https://www.mapr.com/)

The set of algorithms that matter theoretically is different from the ones that matter commercially. Commercial importance often hinges on ease of deployment, robustness against perverse data and conceptual simplicity. Often, even accuracy can be sacrificed against these other goals. Commercial systems also often live in a highly interacting environment so off-line evaluations may have only limited applicability. I will describe several commercially important algorithms such as Thompson sampling (aka Bayesian Bandits), result dithering, on-line clustering and distribution sketches and will explain what makes these algorithms important in industrial settings.

Sorting the wheat from the chaff when learning in the wild: Tiberio Caetano, Chief Scientist, Ambiata (http://www.ambiata.com)

I'll discuss some of my experience helping organisations to demonstrably cause profit gains by leveraging their 1st party data. I'll focus on trying to label things as either important or not (disclaimer: I label the awareness of the fallibility of one's labelling as important). This is a report on incursions into causal inference, optimisation, uncertainty, large-scale machine learning, non-standard learning settings, human psychology and ethics.