Papers/2609.00059
🧪 Test?View on arXiv

DISTAL: Distillation and Self-Supervised Pretraining for Structure-Agnostic Materials Property Prediction

Not specified in the provided content

self-supervised learningmaterials informaticsmultimodalknowledge distillation
2609.00059
Builder Relevance
80%
1h ago

Abstract

DISTAL proposes a dual-prior framework for structure-agnostic materials property prediction that combines self-supervised compositional pretraining with structure-aware knowledge distillation.

Reality Card

Core Claim

DISTAL achieves the strongest overall performance among all evaluated feature combinations, improving over the reference benchmark on 37 out of 39 benchmark tasks.

Method / Result

The best-performing multimodal configuration combines compositional descriptors, pretrained latent features, and distilled structural features.

Limitations

The main limitation is the reliance on a pretrained ALIGNN teacher, which may affect reproducibility if the teacher model is not accessible.

Paper to code

Verified implementation resources so builders can test the paper’s claims instead of stopping at the abstract.

No verified implementation link has been attached yet. AIBuzzHub will keep this panel separate from unverified search results.
← Back to all papers