List your company
ResearchFormerly WizWand

Optimal Transport map estimation on OT map (a) T1(x)=x (d=20) n=m=5000 (test)

0.067MSEHyCNN
-0.455483.071266.59810.12474Apr 29, 2026
Updated 6mo ago

Evaluation results

MethodLinks
HyCNNArchitecture=(48 × 4)2026.040.067
HyCNNArchitecture=(48 × 6)2026.040.076
HyCNNArchitecture=(48 × 2)2026.040.109
ICNNq (Softplus)Architecture=(64 × 2), tau=12026.040.365
ICNNq (Softplus)Architecture=(64 × 4), tau=12026.040.407
ICNNq (Softplus)Architecture=(64 × 2), tau=102026.040.486
ICNNq (Softplus)Architecture=(64 × 4), tau=102026.040.49
ICNNq (Softplus)Architecture=(64 × 6), tau=102026.040.504
ICNNq (Softplus)Architecture=(64 × 2), tau=0.12026.040.567
ICNNq (Softplus)Architecture=(64 × 6), tau=12026.040.57
ICNNq (ReLU)Architecture=(64 × 2)2026.040.801
ICNNq (Softplus)Architecture=(64 × 4), tau=0.12026.040.858
ICNNq (Softplus)Architecture=(64 × 6), tau=0.12026.040.971
ICNNq (ReLU)Architecture=(64 × 4)2026.041.123
ICNNq (ReLU)Architecture=(64 × 6)2026.041.401
ICNN (ReLU)Architecture=(64 × 6)2026.041.965
ICNN (L-ReLU)Architecture=(64 × 6)2026.042.002
ICNN (ReLU)Architecture=(64 × 4)2026.042.222
ICNN (L-ReLU)Architecture=(64 × 4)2026.042.599
ICNN (ReLU)Architecture=(64 × 2)2026.043.221
ICNN (L-ReLU)Architecture=(64 × 2)2026.044.36
MongeGapArchitecture=(64 × 2)2026.044.532
MongeGapArchitecture=(64 × 4)2026.045.296
MongeGapArchitecture=(64 × 6)2026.045.301
Entropic OTepsilon=12026.046.537
Entropic OTepsilon=0.12026.0412.098
Entropic OTepsilon=102026.0413.129