A Tutorial on Deep Latent Variable Models of Natural Language
It provides an educational resource for researchers and practitioners in machine learning and NLP, but is incremental as it synthesizes existing knowledge rather than presenting new findings.
This tutorial addresses the challenges of combining latent variable models with deep learning for natural language, focusing on intractable posterior inference and non-differentiability in backpropagation, and explores variational inference as a solution.
There has been much recent, exciting work on combining the complementary strengths of latent variable models and deep learning. Latent variable modeling makes it easy to explicitly specify model constraints through conditional independence properties, while deep learning makes it possible to parameterize these conditional likelihoods with powerful function approximators. While these "deep latent variable" models provide a rich, flexible framework for modeling many real-world phenomena, difficulties exist: deep parameterizations of conditional likelihoods usually make posterior inference intractable, and latent variable objectives often complicate backpropagation by introducing points of non-differentiability. This tutorial explores these issues in depth through the lens of variational inference.