List your company
ResearchFormerly WizWand

Optimal Transport Map Estimation on OT Map T3(x) n=5000 (d=50) c (test)

3.82MSEICNN (L-ReLU)
-5.16155.46075116.0825176.70425Apr 29, 2026
Updated 6mo ago

Evaluation results

MethodLinks
ICNN (L-ReLU)Architecture (width x depth)=(64 x 2)2026.043.82
HyCNNArchitecture (width x depth)=(48 x 6)2026.0419.771
HyCNNArchitecture (width x depth)=(48 x 4)2026.0419.84
HyCNNArchitecture (width x depth)=(48 x 2)2026.0421.572
ICNNq (Softplus)Architecture (width x depth)=(64 x 2), tau=12026.0424.779
ICNNq (Softplus)Architecture (width x depth)=(64 x 2), tau=102026.0425.55
ICNNq (ReLU)Architecture (width x depth)=(64 x 2)2026.0426.985
ICNNq (Softplus)Architecture (width x depth)=(64 x 4), tau=102026.0427.468
ICNNq (Softplus)Architecture (width x depth)=(64 x 6), tau=102026.0427.948
ICNNq (Softplus)Architecture (width x depth)=(64 x 2), tau=0.12026.0428.272
ICNNq (Softplus)Architecture (width x depth)=(64 x 6), tau=12026.0428.474
ICNNq (Softplus)Architecture (width x depth)=(64 x 4), tau=12026.0430.605
ICNNq (Softplus)Architecture (width x depth)=(64 x 4), tau=0.12026.0444.097
ICNNq (ReLU)Architecture (width x depth)=(64 x 4)2026.0448.32
ICNN (ReLU)Architecture (width x depth)=(64 x 6)2026.0455.948
ICNN (L-ReLU)Architecture (width x depth)=(64 x 6)2026.0459.944
ICNNq (ReLU)Architecture (width x depth)=(64 x 6)2026.0461.714
MongeGapArchitecture (width x depth)=(64 x 6)2026.0465.164
ICNN (ReLU)Architecture (width x depth)=(64 x 4)2026.0466.244
MongeGapArchitecture (width x depth)=(64 x 4)2026.0475.989
MongeGapArchitecture (width x depth)=(64 x 2)2026.0484.647
ICNN (L-ReLU)Architecture (width x depth)=(64 x 4)2026.0485.385
Entropic OTepsilon=102026.0489.193
ICNN (ReLU)Architecture (width x depth)=(64 x 2)2026.0489.585
Entropic OTepsilon=12026.04169.342
Entropic OTepsilon=0.12026.04189.406
ICNNq (Softplus)Architecture (width x depth)=(64 x 6), tau=0.12026.04228.345