Query-Conditioned Evidential Keyframe Sampling for MLLM-Based Long-Form Video Understanding

Query-Conditioned Evidential Keyframe Sampling for MLLM-Based Long-Form Video Understanding explores A principled framework for query-conditioned evidential keyframe sampling optimizes multimodal LLM performance on long-form videos by maximizing conditional mutual information.. Commercial viability score: 7/10 in Video Understanding.