Design Motifs for Probabilistic Generative Design

Geoffrey Roeder, Nathan Killoran, Will Grathwohl, David Duvenaud

Abstract:

Generative models can be used to produce designs that obey hard-to-specify constraints while still producing plausible examples.
Recent examples of this include drug design, text with desired sentiment, or images with desired captions.
However, most previous applications of generative models to design are based on bespoke, ad-hoc procedures.
We give a unifying treatment of generative design based on probabilistic generative models.
Some of these models can be trained end-to-end, can take advantage of both labelled and unlabelled examples, and automatically trade off between different design goals.

Keywords: generative design, latent variable models, inference, generative models, deep learning, Bayesian machine learning

TL;DR: Proposed language of "design motifs" for generative models of structured data, intended for design, exposes strengths of semi-supervised learning and joint training