The Science of Prediction

Artificial Intelligence is ubiquitous: it is used in chat-bots, medical applications, and in games to defeat world champions at Chess, Go, Poker and StarCraft. In this talk, we’ll present our recent work on applying AI to help design and play-test Tabletop Board Games (such as Settlers of Catan, Ticket to Ride or Pandemic), which are intrinsically social games that bring people together around a table. We are working with board game designers to create AI tools they can use to build better tabletop games, so they can make fun and safe social environments for their players.

An important goal in atmospheric science research is to find simplified models of the atmosphere to better understand weather systems, model their evolution over time, and predict future outcomes. Weather data sets present many challenges: they are often large and contain uncertain or missing data. In this talk I will explain how techniques from the field of Topological Data Analysis — a relatively new discipline in data science — allow to study properties of data that one can think of as giving a summary of its "shape", and how these methods help in studying large and noisy weather data sets.