#71 Artificial Intelligence, Deepmind & Social Change, with Julien Cornebise
Learning Bayesian Statistics
Using Stable Diffusion to Generate Full Resolution Images
Stable Diffusion is a generative model for images. Enter your text, describe what you wanted to generate and it draws it in very photorealistic ways. These are technology that even two years ago, where I say, no, there's absolutely no way this can ever be done.
Proudly sponsored by PyMC Labs, the Bayesian Consultancy. Book a call, or get in touch!
This episode will show you different sides of the tech world. The one where you research and apply algorithms, where you get super excited about image recognition and AI-generated art. And the one where you support social change actors — aka the “AI for Good” movement.
My guest for this episode is, quite naturally, Julien Cornebise. Julien is an Honorary Associate Professor at UCL. He was an early researcher at DeepMind where he designed its early algorithms. He then worked as a Director of Research at ElementAI, where he built and led the London office and “AI for Good” unit.
After his theoretical work on Bayesian methods, he had the privilege to work with the NHS to diagnose eye diseases; with Amnesty International to quantify abuse on Twitter and find destroyed villages in Darfur; with Forensic Architecture to identify teargas canisters used against civilians.
Other than that, Julien is an avid reader, and loves dark humor and picking up his son from school at the 'hour of the daddies and the mommies”.
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, Adam Bartonicek, William Benton, James Ahloy, Robin Taylor, Thomas Wiecki, Chad Scherrer, Nathaniel Neitzke, Zwelithini Tunyiswa, Elea McDonnell Feit, Bertrand Wilden, James Thompson, Stephen Oates, Gian Luca Di Tanna, Jack Wells, Matthew Maldonado, Ian Costley, Ally Salim, Larry Gill, Joshua Duncan, Ian Moran, Paul Oreto, Colin Caprani, George Ho, 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, Luis Iberico, Tomáš Frýda, Ryan Wesslen, Andreas Netti, Riley King, Aaron Jones, 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, David Haas, Robert Yolken, Or Duek, Pavel Dusek and Paul Cox.
Visit https://www.patreon.com/learnbayesstats to unlock exclusive Bayesian swag ;)
Links from the show:
- Julien’s website: https://cornebise.com/julien/
- Julien on Twitter: https://twitter.com/JCornebise
- Julien on LinkedIn: https://www.linkedin.com/in/juliencornebise/
- Julien on Scholar: https://scholar.google.co.uk/citations?user=6fkVVz4AAAAJ&hl=en&oi=ao
- Stable Diffusion is a really big deal: https://simonwillison.net/2022/Aug/29/stable-diffusion/
- LBS #21, Gaussian Processes, Bayesian Neural Nets & SIR Models, with Elizaveta Semenova: https://learnbayesstats.com/episode/21-gaussian-processes-bayesian-neural-nets-sir-models-with-elizaveta-semenova/
- pymc.find_constrained_prior function: https://www.pymc.io/projects/docs/en/stable/api/generated/pymc.find_constrained_prior.html#pymc.find_constrained_prior
- LBS #50, Ta(l)king Risks & Embracing Uncertainty, with David Spiegelhalter: https://learnbayesstats.com/episode/50-talking-risks-embracing-uncertainty-david-spiegelhalter/
- LBS #67 Exoplanets, Cool Worlds & Life in the Universe, with David Kipping: https://learnbayesstats.com/episode/67-exoplanets-cool-worlds-life-in-universe-david-kipping/
Abstract
Julien Cornebise goes on a deep dive into deep learning with us in episode 71. He calls himself a “passionate, impact-driven scientist in Machine Learning and Artificial Intelligence”. He holds an Honorary Associate Professor position at UCL, was an early researcher at DeepMind, went on to become Director of Research at ElementAI and worked with institutions ranging from the NHS in Great-Britain to Amnesty International.
He is a strong advocate for using Artificial Intelligence and computer engineering tools for good and cautions us to think carefully about who we develop models and tools for. Ask the question: What could go wrong? How could this be misused? The list of projects where he used his computing skills for good is long and divers: With the NHS he developed methods to measure and diagnose eye diseases. For Amnesty International he helped quantify the abuse female journalists receive on Twitter, based on a database of tweets labeled by volunteers.
Beyond these applied projects, Julien and Alex muse about the future of structured models in times of more and more popular deep learning approaches and the fascinating potential of these new approaches. He advices anyone interested in these topics to be comfortable with experimenting by themselves and potentially breaking things in a non-consequential environment.
And don’t be too intimidated by more seasoned professionals, he adds, because they probably have imposter-syndrome themselves which is a sign of being aware of ones own limitations.
Automated Transcript
Please note that the following transcript was generated automatically and may therefore contain errors. Feel free to reach out if you’re willing to correct them.
Remember Everything You Learn from Podcasts
