Search papers, labs, and topics across Lattice.
Affiliation:
3
0
6
Severity scores are meaningless without spatial grounding: gating RL rewards on strictly ordered localization and classification dependencies forces MLLMs to diagnose visual degradation rather than hallucinate holistic quality metrics.
KBK achieves a 2.7% accuracy boost over previous methods in multi-label class-incremental learning, all without using replay buffers.
OKR achieves a remarkable +6.5% mAP improvement over existing methods by preventing knowledge interference in domain-incremental object detection.
