List your company
ResearchFormerly WizWand

Graph Regression on ZINC (MAE)

0.017MAEESA
0.002680.099340.1960.29266Jun 14, 2020Jun 10, 2021Jun 6, 2022Jun 2, 2023May 28, 2024May 24, 2025May 21, 2026
Updated 5mo ago

Evaluation results

MethodLinks
ESAPE=true2024.020.017
GPS2024.020.024
ESA2024.020.027
Exphormer2024.020.041
TAU-GRITTokenizer=TAU, Backbone=GRIT2025.100.051
TIGT2024.020.057
GRIT2024.020.059
GRIT2025.100.059
GPS + GFSAGFSA=true2023.120.069
QUESTAttention=QUEST2026.030.069
GraphGPS2026.040.069
GPSGFSA=false2023.120.07
GPSruns=4, seeds=42024.100.07
GPS2024.020.07
StandardAttention=Standard2026.030.07
GraphGPS2025.100.07
Cluster-GTruns=4, seeds=42024.100.071
Graph MLP-Mixerruns=4, seeds=42024.100.073
Graph-ViT-MLPMixer2026.030.073
FragNet2026.030.078
HGODE2026.040.078
HGODE2026.040.078
CIN2024.020.079
FROND2026.040.079
GIN-AK+2024.020.08
Graphormer-GD2024.020.081
DRAGON2026.040.081
Subgraphormer2026.040.082
GREAD-BS2026.040.084
Graph-ViT + GFSAGFSA=true2023.120.0845
Graph-ViTGFSA=false2023.120.0849
Graph ViTruns=4, seeds=42024.100.085
CRaW12024.020.085
GCN+2026.040.087
BuNN2026.040.087
GatedGCN+LSPE2024.020.09
K-Subgraph SAT2024.020.094
GSN2024.020.101
Polynormer2024.020.101
Cy2C-GNNs2024.020.102
GREAD-Exp2026.040.102
EGT#Param=≈500K2021.080.108
EGT2024.020.108
EGT2025.100.108
Graphormer + GFSA#Params=500K2023.120.1189
ALL-IN-SPECIALIZEDTraining Scope=Trained per dataset, Node Properties=Standard2026.050.1195
DGN#Param=≈500K2021.080.122
Graphormer#Param=≈500K2021.080.122
Graphormerruns=4, seeds=42024.100.122
Graphormer2024.020.122
ALL-INTraining Scope=Trained on all datasets, Node Properties=Standard2026.050.1237
Graphormer#Params=500K2023.120.124
FLODE2026.040.124
ACMP2026.040.124
Our Method2026.030.131
Graphormer2026.030.132
SAN+LapPEruns=4, seeds=42024.100.139
SAN2024.020.139
PNA#Param=≈500K2021.080.142
EGT#Param=≈100K2021.080.143
HGODEvariant=w/o hysteresis2026.040.145
GRAND2026.040.147
Sheaf Diff.2026.040.148
ALL-IN-SPECIALIZEDTraining Scope=Trained per dataset, Node Properties=0 props2026.050.148
CGNN2026.040.153
ALL-INTraining Scope=Trained on all datasets, Node Properties=0 props2026.050.1557
HGODEvariant=w/o topo. search2026.040.162
DGN#Param=≈100K2021.080.168
DGN2024.020.168
HGODEvariant=w/o force margin2026.040.172
GCNII2026.040.178
PNA#Param=≈100K2021.080.188
PNA2024.020.188
GraphiTruns=4, seeds=42024.100.202
GT#Param=≈500K2021.080.214
GatedGCN#Param=≈500K2021.080.214
PNAedge features=none2022.050.219
GT#Param=≈100K2021.080.226
GTruns=4, seeds=42024.100.226
GraphGT2026.030.226
GDE2026.040.239
LRGA + GatedGCNParameter Budget=446K2020.060.249
InfoGraphPre-training Type=Contrastive2023.090.254
GatedGCNParameter Budget=504K2020.060.282
GatedGCN#Param=≈100K2021.080.282
GatedGCN2024.020.282
tGNNedge features=none2022.050.301
PHM-GNNedge features=none2022.050.32
GHC2024.070.337
GTTok.=RW, Mode=Pad.2026.050.342
HANreproduced=official implementation2026.030.348
F-GINPre-training Type=Supervised (Reference)2023.090.35
LRGA + GatedGCNParameter Budget=94K2020.060.355
GCN2026.040.359
LRGA + GATParameter Budget=536K2020.060.36
GCNParameter Budget=505K2020.060.367
GCN#Param=≈500K2021.080.367
GCN2024.020.367
GatedGCNParameter Budget=105K2020.060.375
GIN#Param=≈500K2021.080.375