Molecular Property Prediction On Esol

المقاييس

RMSE

النتائج

نتائج أداء النماذج المختلفة على هذا المعيار القياسي

اسم النموذج	R2	RMSE	Paper Title	Repository
GAT	0.930±0.007	0.540±0.027	Graph Attention Networks	-
ChemRL-GEM	-	0.798	ChemRL-GEM: Geometry Enhanced Molecular Representation Learning for Property Prediction	-
Graphormer	0.908±0.021	0.618±0.068	Do Transformers Really Perform Bad for Graph Representation?	-
XGBoost	-	0.99	MoleculeNet: A Benchmark for Molecular Machine Learning	-
GIN	0.938±0.011	0.509±0.044	How Powerful are Graph Neural Networks?	-
D-MPNN	-	1.050	Analyzing Learned Molecular Representations for Property Prediction	-
MPNN	-	0.58	MoleculeNet: A Benchmark for Molecular Machine Learning	-
GCN	0.936±0.006	0.520±0.024	Semi-Supervised Classification with Graph Convolutional Networks	-
GATv2	0.928±0.005	0.549±0.020	How Attentive are Graph Attention Networks?	-
S-CGIB	-	0.816±0.019	Pre-training Graph Neural Networks on Molecules by Using Subgraph-Conditioned Graph Information Bottleneck
PNA	0.942±0.006	0.493±0.026	Principal Neighbourhood Aggregation for Graph Nets	-
ChemBFN	-	0.884	A Bayesian Flow Network Framework for Chemistry Tasks	-
DropGIN	0.935±0.012	0.520±0.048	DropGNN: Random Dropouts Increase the Expressiveness of Graph Neural Networks	-
SPMM	-	0.810	Bidirectional Generation of Structure and Properties Through a Single Molecular Foundation Model	-
Uni-Mol	-	0.788	Uni-Mol: A Universal 3D Molecular Representation Learning Framework
SMA	-	0.623	Self-Guided Masked Autoencoders for Domain-Agnostic Self-Supervised Learning	-
ChemBERTa-2 (MTR-77M)	-	0.889	ChemBERTa-2: Towards Chemical Foundation Models	-
TokenGT	0.892±0.036	0.667±0.103	Pure Transformers are Powerful Graph Learners	-
GraphGPS	0.911±0.003	0.613±0.010	Recipe for a General, Powerful, Scalable Graph Transformer	-
ESA (Edge set attention, no positional encodings)	0.944±0.002	0.485±0.009	An end-to-end attention-based approach for learning on graphs	-

0 of 20 row(s) selected.