Offline Multi-Objective Optimization on Off-MOO-Bench Sci-Design benchmark leaderboard · SOTA2 Research

Offline Multi-Objective Optimization on Off-MOO-Bench Sci-Design

1.8Average IGDoffline RankD(best)

Evaluation Results

Method	Links
D(best)2026.06		1.8
ParetoFlow2026.06		3.21
MOBO2026.06		4.95
Multiple Models2026.06		6.1
Multi Head + GradNorm2026.06		6.2
ParetoFlow2026.06		6.21
Multi Head2026.06		6.25
Multiple Models2026.06		6.45
MultipleModels + COMs2026.06		6.72
End-to-End2026.06		6.97
Multi Head2026.06		7.17
Multiple Models2026.06		7.4
End-to-End + PcGrad2026.06		7.47
DOMOO2026.06		7.55
Multiple Models2026.06		7.97
MOBO-qParEGO2026.06		8
Multiple Models + RoMA2026.06		8.15
MOBO2026.06		8.3
Multi Head + PcGrad2026.06		8.35
DOMOO2026.06		8.6
End-to-End2026.06		8.68
Multi Head2026.06		8.95
End-to-End2026.06		8.97
End-to-End2026.06		9.25
Multi Head2026.06		9.25
Multiple Models + ICT2026.06		9.55
End-to-End + GradNorm2026.06		9.57
Multiple Models2026.06		9.57
D(best)2026.06		9.62
Multiple Models + IOM2026.06		9.78
MOBO2026.06		10.27
Multiple Models + Tri-Mentoring2026.06		10.3
Multiple Models2026.06		11.03
Multiple Models2026.06		11.62