AI RESEARCH
NexusFlow: Unifying Disparate Tasks under Partial Supervision via Invertible Flow Networks
arXiv CS.CV
•
ArXi:2512.06251v2 Announce Type: replace Partially Supervised Multi-Task Learning (PS-MTL) aims to leverage knowledge across tasks when annotations are incomplete. Existing approaches, however, have largely focused on the simpler setting of homogeneous, dense prediction tasks, leaving the realistic challenge of learning from structurally diverse tasks unexplored. To this end, we