AI Paper Atlas

Research map / Label-efficient and robust learning / Self-supervised and contrastive

TriForces: Augmenting Atomistic GNNs for Transferable Representations

Ali Ramlaoui, Alexandre Duval, Hannah Bull, Victor Schmidt, Hugues Talbot, Fragkiskos D. Malliaros, Joseph Musielewicz

ICML 2026 · Label-efficient and robust learning › Self-supervised and contrastive · approach: Invariance-based self-supervised learning · 0 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
add composition and structure streams plus SSL pretraining (TriForces)Orb-v3 Conservative · OMat24 valenergy MAE (meV/atom): 107 -> 19.4Improved
add SSL pretraining on top of the three streams (ablation)Orb-v3 Conservative · OMat24 valforce MAE (meV/A): 149 -> 95.5Improved
add three streams without SSL (ablation)Orb-v3 Conservative · OMat24 valforce MAE (meV/A): 150 -> 149No clear effect
add three streams without SSL (ablation)MACE · OMat24 valenergy MAE (meV/atom): 117 -> 81.6Improved
add three streams without SSL (ablation)eSEN · MPtrj subsetforce MAE (meV/A): 75.9 -> 93.3Hurt
add SSL pretraining on top of streams (ablation)Orb-v3 Direct · OMat24 valstress MAE: 6.5 -> 7.8Hurt
add TriForces three-stream plus SSL at 20K training sampleseSEN Direct · OMat24 fine-tuningenergy MAE (meV/atom): 81.3 -> 34.6Improved
add SSL initialization at full OMat24 scale (ablation)TriForces-Streams (eSEN) · Full OMat24energy MAE (meV/atom): 13.7 -> 13.5No clear effect
use SSL instead of supervised pretraining on LeMat-Bulk (ablation)eSEN · OMat24 fine-tuningforce MAE (meV/A): 106.0 -> 86.8Improved
add composition stream to interaction stream (ablation)eSEN-sm conservative · OMat24-2M (no SSL)energy MAE (meV/atom): 102 -> 16Improved
add structure stream to composition+interaction (ablation)eSEN-sm conservative · OMat24-2M (no SSL)crystal-system probe accuracy (%): 50.6 -> 68.1Improved
add TriForces to Orb-v3 on Matbench DiscoveryOrb-v3 · Matbench Discovery (unique prototypes)MAE (eV/atom): 0.024 to 0.020Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy