Poster
Meta-Learning MCMC Proposals
Tongzhou Wang · YI WU · Dave Moore · Stuart Russell

Thu Dec 6th 10:45 AM -- 12:45 PM @ Room 210 #47

Effective implementations of sampling-based probabilistic inference often require manually constructed, model-specific proposals. Inspired by recent progresses in meta-learning for training learning agents that can generalize to unseen environments, we propose a meta-learning approach to building effective and generalizable MCMC proposals. We parametrize the proposal as a neural network to provide fast approximations to block Gibbs conditionals. The learned neural proposals generalize to occurrences of common structural motifs across different models, allowing for the construction of a library of learned inference primitives that can accelerate inference on unseen models with no model-specific training required. We explore several applications including open-universe Gaussian mixture models, in which our learned proposals outperform a hand-tuned sampler, and a real-world named entity recognition task, in which our sampler yields higher final F1 scores than classical single-site Gibbs sampling.

Author Information

Tongzhou Wang (Facebook AI Research)
YI WU (UC Berkeley)
Dave Moore (Google)
Stuart Russell (UC Berkeley)

More from the Same Authors