Offline Multi-Objective Optimization on Off-MOO-Bench MO-NAS benchmark leaderboard · SOTA2 Research

Offline Multi-Objective Optimization on Off-MOO-Bench MO-NAS

3.89Average IGDoffline RankEnd-to-End

Evaluation Results

Method	Links
End-to-End2026.06		3.89
Multi Head2026.06		4
End-to-End2026.06		4.64
Multiple Models2026.06		4.95
Multi Head2026.06		5.01
Multiple Models2026.06		5.28
MOBO2026.06		6.26
End-to-End2026.06		6.54
MultipleModels + COMs2026.06		6.56
MOBO2026.06		6.58
Multiple Models2026.06		6.9
End-to-End + PcGrad2026.06		6.99
Multiple Models + RoMA2026.06		7.12
DOMOO2026.06		7.14
Multiple Models2026.06		7.3
Multi Head + PcGrad2026.06		7.42
Multi Head2026.06		7.53
Multiple Models2026.06		7.68
Multiple Models + IOM2026.06		7.85
DOMOO2026.06		8.7
Multiple Models + ICT2026.06		8.8
Multiple Models2026.06		9.11
ParetoFlow2026.06		9.19
D(best)2026.06		9.31
Multiple Models2026.06		9.52
ParetoFlow2026.06		9.6
Multiple Models + Tri-Mentoring2026.06		9.71
End-to-End2026.06		10.93
MOBO-qParEGO2026.06		11.27
End-to-End + GradNorm2026.06		12.1
MOBO2026.06		12.15
Multi Head + GradNorm2026.06		12.58
Multi Head2026.06		12.68
D(best)2026.06		12.83