Amortized Bayesian inference for clustering models
This addresses efficiency issues in Bayesian clustering for researchers and practitioners, though it appears incremental as an optimization of existing inference methods.
The paper tackles the computational challenge of Bayesian inference for clustering models by developing an amortized method that maps symmetry-invariant representations to conditional probabilities, achieving iid samples from the approximate posterior with the same cost as a single Gibbs sweep and applicability to both conjugate and non-conjugate models.
We develop methods for efficient amortized approximate Bayesian inference over posterior distributions of probabilistic clustering models, such as Dirichlet process mixture models. The approach is based on mapping distributed, symmetry-invariant representations of cluster arrangements into conditional probabilities. The method parallelizes easily, yields iid samples from the approximate posterior of cluster assignments with the same computational cost of a single Gibbs sampler sweep, and can easily be applied to both conjugate and non-conjugate models, as training only requires samples from the generative model.