TGViewer
Science Scroll Science Scroll @sciencescroll · 382 subscribers
Post #76184 12
📄 Less can be more for predicting properties with large language models

📝 مدل‌های زبانی بزرگ در یادگیری اطلاعات مختصاتی داده‌های مختصات-دسته ناتوان‌اند و برای پیش‌بینی خواص مواد، معماری‌های هندسی تخصصی بهتر عمل می‌کنند.

📚 Journal: Nature Computational Science #NatureComputationalScience
📅 Date: 2026-10-12
🔬 Research Article | 🔓 Open Access | 📥 PDF
🔗 10.1038/s43588-026-01045-2

🏷 large language models · property prediction · coordinate-category data · geometric architectures · MatText
More from @sciencescroll
  1. Oct 13, 2026📄 Stable deep-blue organic light-emitting diodes with reduced efficiency roll-off via acc…
  2. Oct 13, 2026📄 Solid-state emissive acceptors enable thermally accessible radiative pathways for effic…
  3. Oct 13, 2026📄 Cost effectiveness of synthetic fuel and CO2 removal for net-zero aviation 📝 این مقاله…
  4. Oct 13, 2026📄 Advancing Sulfide Electrolyte/Silicon-Based Anode Systems in All-Solid-State Batteries:…
  5. Oct 13, 2026📄 Molecular weaving goes multilevel 📝 این مقاله چشم‌اندازهایی درباره بافتن مولکولی چندسط…
  6. Oct 13, 2026📄 A review on regulation of non-equilibrium solidification microstructure in selective la…
Threads Profile ViewerView any public Threads profile without an account.Open ThreadLook →Writing with AI? Make it sound human.Metric37 rewrites AI drafts so they read naturally. Free AI detector, 1,500 words free.Try Metric37 →