Wed, 12/18/2019

Wuejie's first-author paper on the use of unsupervised and semi-supervised VAEs to learn optimal coarse-grained representations _and_ neural network interatomic potentials _at the same time_ is now published. There is an...

Mon, 10/07/2019

Daniel and Rafa's article, in collaboration with the Olivetti group at MIT DMSE, was published today in Nature Materials! The work combines graph theory, crystallography and extensive literature search to quantify, explain and predict phase transformations...

Wed, 07/10/2019

Daniel has recently posted a preprint for his upcoming book chapter "Generative Models for Automatic Chemical Design" where he reviews some classics and dives deep into the recent applications of neural networks to the inverse design problem...

Mon, 07/01/2019

We are grateful to the MITEI sponsors for the support of this avenue of work, and to Kathryn Luu for helping us tell this story. Check out our preprint for some preliminary results in the area.

Thu, 06/13/2019

Yesterday, the team took a few hours away from computers to assemble new furniture for everyone's offices. Simulation folks hard at work putting together standing desks. Looking forward to that ergonomic feeling.


Sun, 05/19/2019

Google AI has generously awarded us a Faculty Research Award to support our work in coarse-graining and inverse design of soft matter! We are looking forward to collaborating and learning from our partners at Google!

Wed, 04/24/2019

Congratulations to James on being awarded a National Defense Science and Engineering Graduate fellowship! With this support he can continue to do great work on dimensionality reduction in materials problems.


Thu, 08/23/2018

Some of the science from the last year is now seeing the light of day. In a recent pre-print,...

Sat, 06/23/2018

Rafa's first paper after joining MIT is out: An opinion piece about the near future of machine learning in materials discovery.

Wed, 05/23/2018

Congratulations team! The group has received a MIT Energy Initiative award to pursue Deep learning of contracted basis sets for rapid quantum calculation of thermochemistry and other energy processes. These tools will allow theoretical chemists to explore reaction energies and rates...