Learning Bayesian Statistics #86 Exploring Research Synchronous Languages & Hybrid Systems, with Guillaume Baudart
Proudly sponsored by PyMC Labs, the Bayesian Consultancy. Book a call, or get in touch!
This episode is unlike anything I’ve covered so far on the show. Let me ask you: Do you know what a research synchronous language is? What about hybrid systems? Last try: have you heard of Zelus, or ProbZelus?
If you answered “no” to one of the above, then you’re just like me! And that’s why I invited Guillaume Baudart for this episode — to teach us about all these fascinating topics!
A researcher in the PARKAS team of Inria, Guillaume's research focuses on probabilistic and reactive programming languages. In particular, he works on ProbZelus, a probabilistic extension to Zelus, itself a research synchronous language to implement hybrid systems.
To simplify, Zelus is a modeling framework to simulate the dynamics of systems both smooth and subject to discrete dynamics — if you’ve ever worked with ODEs, you may be familiar with these terms.
If you’re not — great, Guillaume will explain everything in the episode! And I know it might sound niche, but this kind of approach actually has very important applications — such as proving that there are no bugs in a program.
Guillaume did his PhD at École Normale Supérieure, in Paris, working on reactive programming languages and quasi-periodic systems. He then worked in the AI programming team of IBM Research, before coming back to the École Normale Supérieure, working mostly on reactive and probabilistic programming.
In his free time, Guillaume loves spending time with his family, playing the violin with friends, and… cooking!
Our theme music is « Good Bayesian », by Baba Brinkman (feat MC Lars and Mega Ran). Check out his awesome work at https://bababrinkman.com/ !
Thank you to my Patrons for making this episode possible!
Yusuke Saito, Avi Bryant, Ero Carrera, Giuliano Cruz, Tim Gasser, James Wade, Tradd Salvo, William Benton, James Ahloy, Robin Taylor,, Chad Scherrer, Nathaniel Neitzke, Zwelithini Tunyiswa, Bertrand Wilden, James Thompson, Stephen Oates, Gian Luca Di Tanna, Jack Wells, Matthew Maldonado, Ian Costley, Ally Salim, Larry Gill, Ian Moran, Paul Oreto, Colin Caprani, Colin Carroll, Nathaniel Burbank, Michael Osthege, Rémi Louf, Clive Edelsten, Henri Wallen, Hugo Botha, Vinh Nguyen, Raul Maldonado, Marcin Elantkowski, Adam C. Smith, Will Kurt, Andrew Moskowitz, Hector Munoz, Marco Gorelli, Simon Kessell, Bradley Rode, Patrick Kelley, Rick Anderson, Casper de Bruin, Philippe Labonde, Michael Hankin, Cameron Smith, Tomáš Frýda, Ryan Wesslen, Andreas Netti, Riley King, Yoshiyuki Hamajima, Sven De Maeyer, Michael DeCrescenzo, Fergal M, Mason Yahr, Naoya Kanai, Steven Rowland, Aubrey Clayton, Jeannine Sue, Omri Har Shemesh, Scott Anthony Robson, Robert Yolken, Or Duek, Pavel Dusek, Paul Cox, Trey Causey, Andreas Kröpelin, Raphaël R, Nicolas Rode, Gabriel Stechschulte, Arkady, Kurt TeKolste, Gergely Juhasz, Marcus Nölke, Maggi Mackintosh, Grant Pezzolesi, Avram Aelony and Joshua Meehl.
Visit https://www.patreon.com/learnbayesstats to unlock exclusive Bayesian swag ;)
Links from the show:
- Guillaume’s website: https://guillaume.baudart.eu/
- ProbZelus on GitHub: https://github.com/IBM/probzelus
- Zelus docs: https://zelus.di.ens.fr/
- Short Zelus introduction: https://www.di.ens.fr/~pouzet/bib/hscc13.pdf
- Guillaume’s course : https://wikimpri.dptinfo.ens-cachan.fr/doku.php?id=cours:c-2-40
- LBS #74 – Optimizing NUTS and Developing the ZeroSumNormal Distribution, with Adrian Seyboldt: https://learnbayesstats.com/episode/74-optimizing-nuts-developing-zerosumnormal-distribution-adrian-seyboldt/
- ProbZelus (design, semantics, delayed-sampling): https://dl.acm.org/doi/abs/10.1145/3385412.3386009
- Semi-symbolic inference: https://dl.acm.org/doi/abs/10.1145/3563347
- Static analysis for bounded memory inference: https://dl.acm.org/doi/abs/10.1145/3485492
Abstract
Guillaume Baudart is researcher at Inria in the PARKAS team at the Département d'Informatique (DI) of the École normale supérieure. He joins us for episode 86 to tell us about ProbZelus, a synchronous probabilistic programming language, that he develops.
We have not covered synchronous languages yet, so, Guillaume gives us some context on this kind of programming approach and how ProbZelus adds probabilistic notions to it.
He explains the advantages of the probabilistic aspects of ProbZelus and what practitioners may profit from it.
For example, synchronous languages are used to program and test autopilots of planes and ensure that they do not have any bugs. ProbZelus may be useful here as Guillaume argues.
Finally, we also touch upon his teaching work and what difficulties he encounters in teaching probabilistic programming.
Transcript
Please note that this is an automated transcript that may contain errors. Feel free to reach out if you're willing to correct them.