Training energy-based variational autoencoders

A computer-implemented method for creating a generative model, the method comprising:

• performing one or more operations based on a training dataset to generate a trained generative model, wherein the trained generative model includes a first component that converts a plurality of data points included in the training dataset into a set of latent variable values, a second component that learns a distribution of the set of latent variable values across the training dataset, and a third component that converts one or more latent variable values into one or more output distributions;
• performing one or more operations to train an energy-based model to learn an energy function based on a first set of values sampled from one or more first distributions associated with the training dataset and a second set of values sampled from one or more second distributions used during operation of one or more portions of the trained generative model;
• and
• creating a joint model that includes the one or more portions of the trained generative model and the energy-based model, wherein the joint model applies one or more energy values generated via the energy-based model to a third set of values sampled from the one or more second distributions to produce a fourth set of values,
• wherein, in operation, the joint model produces the fourth set of values in order to generate a new data point that is not included in the training dataset.

One or more non-transitory computer readable media storing instructions that, when executed by one or more processors, cause the one or more processors to perform the steps of:

• performing one or more operations based on a training dataset to generate a trained generative model, wherein the trained generative model includes a first component that converts a plurality of data points included in the training dataset into a set of latent variable values, a second component that learns a distribution of the set of latent variable values across the training dataset, and a third component that converts one or more latent variable values into one or more output distributions;
• performing one or more operations to train an energy-based model to learn an energy function based on a first set of values sampled from one or more first distributions associated with the training dataset and a second set of values sampled from one or more second distributions used during operation of one or more portions of the trained generative model;
• and
• creating a joint model that includes the one or more portions of the trained generative model and the energy-based model, wherein the joint model applies one or more energy values generated via the energy-based model to a third set of values sampled from the one or more second distributions to produce a fourth set of values,
• wherein, in operation, the joint model produces the fourth set of values in order to generate a new data point that is not included in the training dataset.