Contrastive KERMT pretraining on molecular graphs yields 7.6-9.9% average gains over KERMT baseline on Biogen, ExpansionRX, and ChEMBL-MT ADME endpoints via a single probabilistic latent-variable objective and task-specific GNN heads.
arXiv preprint arXiv:1910.04153 , year=
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
-
Probabilistic Contrastive Pretraining for Multi-task ADME Property Prediction
Contrastive KERMT pretraining on molecular graphs yields 7.6-9.9% average gains over KERMT baseline on Biogen, ExpansionRX, and ChEMBL-MT ADME endpoints via a single probabilistic latent-variable objective and task-specific GNN heads.