JointFM is a zero-shot foundation model for joint distributional prediction of coupled time series, trained on synthetic SDEs and reducing energy loss by 21.1% versus the strongest baseline on unseen synthetic oracles.
Title resolution pending
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
JointFM-0.1: A Foundation Model for Multi-Target Joint Distributional Prediction
JointFM is a zero-shot foundation model for joint distributional prediction of coupled time series, trained on synthetic SDEs and reducing energy loss by 21.1% versus the strongest baseline on unseen synthetic oracles.