Deep Ensembles as Approximate Bayesian Inference | Deep Ensembles as Approximate Bayesian Inference

Deep Ensembles as Approximate Bayesian Inference Andrew Gordon Wilson, Pavel Izmailov We will describe how deep ensembles approximate the Bayesian posterior predictive distribution

Deep Ensembles as Approximate Bayesian Inference Andrew Gordon Wilson, Pavel Izmailov October 21, 2021 Science is about updating our beliefs based on new evidence. It sounds simple and obvious, but this task is often extraordinarily difficult in practice. In reality, it’s easy to get invested in a particular narrative. How often have you encountered reviewers who refuse to update their assessments, even when their criticisms are fully addressed, or shown to be misguided? In those situations, how often does it feel like the goal posts are being moved around, just to maintain the same underlying

related reading

• Bayesian Neural Networkscs.toronto.edu
• A Comprehensive Introduction to Bayesian Deep Learning - Joris Baanjorisbaan.nl
• The Little Book of Deep Learningfleuret.org
• Maybe I was too harsh on deep learning theory (three days ago) — LessWronglesswrong.com
• arxiv.org/pdf/1805.08522arxiv.org
• Writing - betanalpha.github.iobetanalpha.github.io
• Bayesian programming - Wikipediaen.wikipedia.org
• Learning to be Bayesian without Supervisionpapers.nips.cc
• Ensemble learning - Wikipediaen.wikipedia.org
• Bayesian network - Wikipediaen.wikipedia.org
• Teaching LLMs to reason like Bayesiansresearch.google
• David Duvenaudcs.toronto.edu