ReCoGen outperforms six baselines in downstream utility on all sixteen settings across three clinical datasets by decoupling condition representation (per-modality masked autoencoders) from generation (flow matching).
Scientific Data , volume=
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.LG 1years
2026 1verdicts
CONDITIONAL 1representative citing papers
citing papers explorer
-
Represent, Then Generate: Multimodal-Conditioned Time-Series Generation under Irregular Missingness
ReCoGen outperforms six baselines in downstream utility on all sixteen settings across three clinical datasets by decoupling condition representation (per-modality masked autoencoders) from generation (flow matching).