The 8088 The 8088 ← All news
arXiv cs.LG AI Research Apr 21

How to Approximate Inference with Subtractive Mixture Models

★★★★★ significance 2/5

This paper explores the use of Subtractive Mixture Models (SMMs) as a more expressive alternative to classical mixture models for approximate inference. The researchers design new expectation estimators for importance sampling and learning schemes for variational inference to overcome the lack of latent variable semantics in SMMs.

Why it matters Subtractive mixture models offer a more expressive mathematical framework for overcoming traditional limitations in variational inference and importance sampling.
Read the original at arXiv cs.LG

Tags

#inference #mixture models #variational inference #importance sampling

Related coverage