AI RESEARCH
RoAD Benchmark: How LiDAR Models Fail under Coupled Domain Shifts and Label Evolution
arXiv CS.AI
•
ArXi:2601.07855v2 Announce Type: replace-cross For 3D perception systems to operate reliably in real-world environments, they must remain robust to evolving sensor characteristics and changes in object taxonomies. However, existing adaptive learning paradigms struggle in LiDAR settings where domain shifts and label-space evolution occur simultaneously. We