Share&Code: Applied Measure Theory for Probabilistic Modeling

Applied Measure Theory for Probabilistic Modeling
▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔▔
I am glad to welcome Chad Scherrer as our speaker. He build MeasureTheory.jl and will give us an introduction to the ecosystem around it.
The level will be intermediate.

┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉AGENDA┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉
🕡 18:30 - 19:30: Chad Scherrer. Applied Measure Theory for Probabilistic Modeling

Statistics is often framed in terms of probability distributions, but many common operations take us outside of this class of objects. For example, Bayesian inference often leaves us with an unnormalized posterior, which (until normalized) is not a distribution at all, but a measure.

In this talk, we'll motivate the use to measures as a better class of objects for probabilistic computation. The MeasureTheory.jl ecosystem takes a principled approach to design, identifying primitives (measures like Lebesgue and Counting measure that cannot be described in terms of other measures), and a rich set of combinators for building new measures from existing ones. This approach gives good performance and makes it easy to describe measures and related structures (kernels, likelihoods, etc) that are awkward or impossible to express in other systems.

🕢 19:30 - open end: Time for further discussion
┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉┉

The event is going to be online. The link will be made available 10 min before the actual start. Please join 5 min ahead, the session is going to start on-time,
🗓️ Monday 3. July 2023 🕡 CEST 18:30.

Looking forward to see you all,
best,
Stephan Sahm

sponsored by Jolin.io
www.jolin.io