Next session

Angular relational knowledge distillation of machine learning interatomic potentials for scalable catalyst exploration

Wednesday, June 17th, 2026 · 08:30 AM PT / 11:30 AM ET / 5:30 PM CEST

Dr. Seokhyun Choung talk banner
Machine learning interatomic potentials keep getting more accurate, but also far heavier, trading speed for gains in accuracy. This talk introduces Angular Relational Knowledge (ARK) distillation, which shrinks large MLIPs up to ~12x while keeping their accuracy across domains (OMat24, OC20, and SPICE) and architectures (MACE, GemNet, and EquiformerV2), enabling efficient catalyst discovery at scale.

Ongoing Reading Group

LeMaterial Reading Group

The LeMaterial Reading Group is a recurring gathering where we discuss recent papers at the intersection of AI and materials science. Sessions are co-hosted by Entalpic and The AI Alliance.

We cover foundation models, chemistry ML, representation learning, and benchmarks.

Need a one-off invite instead? Use Add to Calendar in the next session card.

Open to everyone. Students, researchers, engineers, and anyone curious about AI for materials science is welcome to join.

Typical schedule

Times vary based on papers. Details are posted in #general.

PT 8:30 AM ET 11:30 AM CET 5:30 PM

Past recordings

Cannot make it live? All sessions are recorded and uploaded to the YouTube playlist.

Suggest a paper

Have an interesting paper you would like us to discuss? We welcome suggestions from all research areas at the intersection of AI and materials science.

  • Email the organizers: luis.pinto [at] entalpic [dot] ai, ali.ramlaoui [at] entalpic [dot] ai.
  • Or share it in the #general channel on Slack.