Automat uses an LLM coding agent to design task-specific compositional descriptors that outperform standard baselines for band gap and Curie temperature prediction in inorganic materials.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cond-mat.mtrl-sci 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Agentic Design of Compositional Descriptors via Autoresearch for Materials Science Applications
Automat uses an LLM coding agent to design task-specific compositional descriptors that outperform standard baselines for band gap and Curie temperature prediction in inorganic materials.