Improving crystal material property prediction with multi-view geometric graph transformer
Liang Zhang, Ziyue Wang, Xin Wang, Xinyu Zhang, Yu Mao, Kong Chen, Yuen Wu · 2026
Read Improving crystal material property prediction with multi-view geometric graph transformer by Liang Zhang, Ziyue Wang, Xin Wang, Xinyu Zhang, Yu Mao, Kong Chen, Yuen Wu (2026) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/pmc-13396671.txt.
Description
Liang Zhang and others lift crystal trait guesses with a multi-view shape graph net in a 2026 study. True crystal shape and web reads are key to machine craft in big crystal sims, yet most nets catch just part of that shape.
Their MGT net learns still scalar reads and turn-aware direction reads at once, fused by a pro-router that weighs each view per task. Multi-task self-taught pretraining cut mean miss by up to 14 percent on past best crystal tests. Stuff labs gain a truer crystal read for trait guesses.
Catalog record
| CREATOR | Liang Zhang, Ziyue Wang, Xin Wang, Xinyu Zhang, Yu Mao, Kong Chen, Yuen Wu |
|---|---|
| YEAR | 2026 |
| COLLECTION | PAPERS |
| FORMAT | TXT |
| FILE SIZE | 84 KB |
| VERSION | 1.0 |
| LICENSE | CC BY 4.0 |
| RIGHTS | CC BY |
| SOURCE | Europe PMC |