Fine-tuning with just 3 data points cuts thermal conductivity error from 47% to 2%
bravo_abad · x · 2026-09-06
Póta et al. argue that scientific foundation models should be judged not by zero-shot accuracy but by how much expensive high-quality data they amortize.
- A pretrained atomistic foundation model serves as a reusable prior; only material-specific parts are fine-tuned, with the learned potential supplying interatomic forces while the Wigner transport equation keeps heat-transfer physics explicit
- Thermal conductivity is a demanding test since it hinges on subtle anharmonic atomic interactions
- For LiBr, adaptation needed just 3 new DFT configurations, dropping the thermal-conductivity error from 47% to 2%
Takeaway: the real test of a scientific foundation model may be how little new data it needs to become trustworthy on a new problem.
More from Research
- ETH Study: Writing Skills and CS Achievement Predict Vibe Coding Proficiency — soumitrashukla9 · 2026-09-06
- Researcher says GPT-6 Astra cracked a month-long theorem overnight and quietly kept working — morqon · 2026-09-06
- Physical AI researcher: even 99% generalization cuts per-task cost by only 50% — m_wulfmeier · 2026-09-06
- 8-Year Study Finds LLMs Have Implicit Symbolic Structure, Multiplicative Codes Enable Generalization — jm_alexia · 2026-09-06
- PaperCompiler compiles papers into file-level specs to fix lossy paper-to-code agents — mohitban47 · 2026-09-06
- MIT MoML conference calls for 2-4 page papers by Sept 11, $1,526 best-paper prize — HannesStaerk · 2026-09-06