Architecture

Scaling Biomolecular Simulations with Equivariant Models

First page

Paper summary

A framework for large-scale biomolecular simulation using equivariant deep learning.

Ask this paper

Key points

Equivariant network scaling: Achieves high accuracy through equivariant deep learning that respects molecular symmetries.

44M atom HIV capsid: Simulated a complete, all-atom, explicitly solvated HIV capsid structure of 44 million atoms.

Nanosecond-scale stable dynamics: Performs nanoseconds-long stable simulations of protein dynamics - much longer than prior ML-MD simulations.

Perlmutter deployment: Scales to the Perlmutter supercomputer, demonstrating ML-accelerated molecular dynamics at HPC scale.

Every Monday