What it does This repo holds the Jupyter notebooks from DeepBayes 2018, a summer school on Bayesian methods in deep learning held in Moscow. The materials cover variational inference and related topics, packaged with a Binder link so you can still execute them without wrestling with six-year-old dependency hell.

The interesting bit Academic course repos usually rot fast; the Binder badge here is a small act of archival defiance. It also captures a specific pedagogical moment—how Bayesian neural networks were taught before the field narrowed toward point estimates at scale.

Key highlights

• 1,045 stars suggest these notebooks found an audience well beyond the 2018 attendees
• Topics tagged: bayesian, deep-learning, variational-inference
• Runnable via mybinder.org without local setup
• Hosted by the Bayesian Methods Research Group (BayesGroup), HSE Moscow
• Companion site at deepbayes.ru (status unknown; sources don’t clarify)

Caveats

• README is minimal: no syllabus, no instructor list, no dependency manifest beyond what Binder infers
• The deepbayes.ru link’s current state is unclear from the sources
• Six years old in a fast-moving field; some techniques may feel dated

Verdict Worth a look if you’re teaching or learning variational inference and want battle-tested notebook examples. Skip if you need state-of-the-art implementations or a structured curriculum with solutions.