Reasoning as Intersection: Consensus-Frame Alignment for Visual Focus in Video-MLLMs

Researchers introduced Consensus Frame GRPO (CF-GRPO), a process-level reward framework for video multimodal large language models that uses consensus frame alignment to improve evidence-aware reasoning without temporal annotations. The method, detailed in arXiv:2606.18441v1, achieved competitive performance on complex video reasoning benchmarks and is available on GitHub.

arXiv:2606.18441v1 Announce Type: new Abstract: Reinforcement learning has improved the reasoning ability of large language models, but applying outcome-only rewards to video multimodal large language models Video-MLLMs provides limited guidance on which visual evidence should support the answer. Inspired by multisensory integration, where consistent cues can enhance the salience and reliability of perceptual estimates, we introduce Consensus Frame GRPO CF-GRPO , a temporal-annotation-free process-level reward framework for evidence-aware video reasoning. CF-GRPO constructs a consensus frame prior from intrinsic video cues, including temporal coverage, scene-transition cues, and query-conditioned visual relevance. It then computes a model-side frame-use score from visual and response representations and optimizes their agreement through the Consensus Frame Reward CFR . With salience-aware sparse aggregation and distribution sharpening, CFR provides a high-contrast reward signal without requiring human temporal annotations. Experiments show that VideoCFR achieves competitive performance across complex video reasoning benchmarks and improves several metrics over representative Video-MLLM and RL baselines, while the consensus prior provides an interpretable view of the evidence frames emphasized during training. The implementation is available at https://github.com/1Pansy/VideoCFR.