Reposted by Arik Reuter
Our ICML'26 "Bayesian Two Cultures" Take on BNNs:
1/ Every 2nd paper: "MCMC for BNNs is infeasible. We thus propose [heuristic uncertainty method]." --- This is a misconception ❌
arxiv.org
Arik Reuter
@arikreuter.bsky.social
99 followers 172 following 28 posts
University of Cambridge and Max Planck Institute for Intelligent Systems I'm interested in amortized inference/PFNs/in-context learning for challenging probabilistic and causal problems. arikreuter.github.io
