Spotlight Publication: “E(n) Equivariant Graph Neural Network for Learning Interactional Properties of Molecules” published in ACS Physical Chemistry B

Congratulations to Kieran Nehil-Puleo! Kieran’s article “E(n) Equivariant Graph Neural Network for Learning Interactional Properties of Molecules” has been selected as a VINSE spotlight publication.

Kieran Nehil-Puleo is a fourth-year graduate student pursuing a Ph.D. in Interdisciplinary Material Science under the guidance of Dr. Peter Cummings and Dr. Jhongyue Yang. This work was achieved by members of the Cummings group. Dr. Cummings’ research group specializes in molecular simulations of fluids and soft matter systems. In a recent publication in ACS Physical Chemistry B, Kieran presents a physics-informed neural network model specializing in the prediction of properties that result from an unknown interaction of molecules in 3D space. In addition, the authors developed a benchmarking dataset consisting of ~10,000 simulated alkylsilane monolayer tribological interactions calculated using non-equilibirum molecular dynamics simulations. The authors demonstrated this type of model’s improved performance resulting from constraining the functions learned to be equally varying to the euclidean group. This alteration reduced the amount of data needed to train accurate property prediction models.

Read Article