PGFM pre-trains an LSTM on physics-simulated lake data and fine-tunes it with conservation-law penalties, improving prediction of lake temperature and dissolved oxygen across 117 lakes.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2025 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Physics-Guided Foundation Model for Scientific Discovery: An Application to Aquatic Science
PGFM pre-trains an LSTM on physics-simulated lake data and fine-tunes it with conservation-law penalties, improving prediction of lake temperature and dissolved oxygen across 117 lakes.