We are interested in discovering user groups from collabo-rative rating datasets of the form i, u, s, where i ∈ I, u ∈ U, and s is the integer rating that user u has assigned to item i. Each user has a set of attributes that help find labeled groups such as young computer scientists in France and American female designers. We formalize the problem of finding user groups whose quality is optimized in multiple dimensions and show that it is NP-Complete. We develop α-MOMRI, an α-approximation algorithm, and h-MOMRI, a heuristic-based algorithm , for multi-objective optimization to find high quality groups. Our extensive experiments on real datasets from the social Web examine the performance of our algorithms and report cases where α-MOMRI and h-MOMRI are useful.