Kavli Affiliate: Wei Gao| Summary:Long-video reasoning is fundamentally constrained by how models acquire and utilize visual evidence. Existing tool-augmented video frameworks often interleave temporal grounding and answer reasoning within a single trajectory, causing early semantic hypotheses to bias evidence localization. We term this failure mode premature semantic commitment, where biased grounding retrieves incomplete evidence and […]
Continue.. EFlow: Learning Evidence Flow for Long-Video Reasoning with Adaptive Reflection