Journal of Cheminformatics

Table 3 Performance (RMSE) of pairing strategies

From: Similarity-based pairing improves efficiency of siamese neural networks for regression tasks and uncertainty quantification

Model	Lipophilicity	Freesolv	ESOL
MLP	0.75	1.60	0.84
MLP-ΔFP	0.74^a (0.77)^b	1.60^a (1.66)^b	0.81^a (0.87)^b
MLP-SNN	0.72^a (0.72)^b	1.50^a (1.57)^b	0.79^a (0.85)^b
RF-FP	0.77	1.91	0.92
RF-ΔFP	0.74^a	1.62^a	0.83^a
Chemformer	0.58	1.07	0.58
Chemformer-SNN	0.61^a (0.75)^c	1.11^a (1.12)^c	0.79^a (0.93)^c
MolBERT [39]	0.60	1.52	0.55

^a Trained with the similarity-based pairing
^b Trained with exhaustive pairing
^c Trained with the random paring where each compound was paired with 50 randomly chosen compounds as an approximation of exhaustive pairing

Back to article page

ISSN: 1758-2946

Contact us

Submission enquiries: journalsubmissions@springernature.com