GENERATIVE MODELS

SYNTHESIS PLANNING CVAES

Conditional variational autoencoders designed to generate action sequences and precursors needed to synthesize a given target material.

From Kim et al. Inorganic Materials Synthesis Planning with Literature-Trained Neural Networks. JCIM. (2020).

OSDA GENERATOR

Generative model that predicts potential organic structure directing agents (OSDA) for a given zeolite structure and chemistry.
Picture from Daeyaert and Deem. Design of organic structure directing agents for polymorph A zeolite beta. JMCA. (2019).

 

MIT, Cambridge, MA, USA

©2020 BY THE SYNTHESIS PROJECT AND THE OLIVETTI GROUP AT MIT.