Diversity-Driven Offline Multi-Objective Optimization via Nested Pareto Set Learning

In three linesDOMOO, an offline multi-objective optimization method, addresses out-of-distribution (OOD) issues by combining cumulative risk control and nested Pareto set learning. Introduces IGD_offline, a tailored indicator for offline settings, to select diverse and convergent solutions.

Summary generated by Claude — human-verified