Timezone: »

The Thermodynamic Variational Objective
Vaden Masrani · Tuan Anh Le · Frank Wood

Wed Dec 11 05:00 PM -- 07:00 PM (PST) @ East Exhibition Hall B + C #194

We introduce the thermodynamic variational objective (TVO) for learning in both continuous and discrete deep generative models. The TVO arises from a key connection between variational inference and thermodynamic integration that results in a tighter lower bound to the log marginal likelihood than the standard variational evidence lower bound (ELBO) while remaining as broadly applicable. We provide a computationally efficient gradient estimator for the TVO that applies to continuous, discrete, and non-reparameterizable distributions and show that the objective functions used in variational inference, variational autoencoders, wake sleep, and inference compilation are all special cases of the TVO. We use the TVO to learn both discrete and continuous deep generative models and empirically demonstrate state of the art model and inference network learning.

Author Information

Vaden Masrani (University of British Columbia)
Tuan Anh Le (MIT)
Frank Wood (University of British Columbia)

More from the Same Authors