Past Meetup

Space Leaks and Gaussian Processes

This Meetup is past

100 people went

Details

Update: unfortunately Noel had to cancel due to travel and we will reschedule his talk. Instead we will have a Haskell themed talk.

Skillsmatter link:

https://skillsmatter.com/meetups/7904-probabilistic-programming-and-gaussian-processes

See you there!

-Dirk, Blair

• Open the cargo bay doors, Haskell: space leaks and laziness - Joe Nash

Haskell provides a lot of abstractions that make it a very expressive, and very safe, programming language. But some of these abstractions can come back to haunt us. In this talk, we'll investigate some neat features of the purely functional, statically typed language, and performance implications of laziness, immutability and purity"

Bio: Joe is a developer advocate at Improbable, helping developers build massive worlds with SpatialOS. An unrepentant Haskell addict, Joe speaks on topics of FP, distributed systems, and programming language theory. A strong believer in the benefits of hackathons, Joe is resident community lead of Hackathon Hackers EU.

• An introduction to modelling uncertainty with Gaussian Processes - Karl Surmacz

Modelling uncertainty allows us to quantify risk, optimize robustly, and ultimately make better-informed decisions. In this talk I will motivate this approach and introduce the concept of Gaussian processes for regression and classification. I will show some practical applications of Gaussian processes, and also present ideas on how to get around one of the big limitations of such statistical models: scalability.

Bio: I am a Principal Data Scientist at McLaren Applied Technologies, where my main interests are in statistical modelling and inference for industrial applications. My previous career has taken me through quantum computing, F1 race strategy and control engineering.