A new high-temperature AIMD benchmark for nine MOFs shows that top uMLIPs like ORB-v3 and fairchem OMAT still produce substantial errors in long-timescale dynamics despite lower static losses.
Wood, Misko Dzamba, Meng Gao, Ammar Rizvi, C
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
-
Benchmarking Universal Machine-Learned Interatomic Potentials for High-Temperature Metal-Organic Framework Chemistry
A new high-temperature AIMD benchmark for nine MOFs shows that top uMLIPs like ORB-v3 and fairchem OMAT still produce substantial errors in long-timescale dynamics despite lower static losses.