CADER: Confidence-Aware Dynamic Evidence Reasoning for Long-Video Understanding

Researchers propose CADER (Confidence-Aware Dynamic Evidence Reasoning), a training-free framework for adaptive long-video reasoning that uses a logit-margin signal to let high-confidence examples exit early and activates a tool-augmented loop for uncertain cases. Experiments on multiple VideoQA benchmarks show CADER improves long-video reasoning while bypassing Stage 2 for high-confidence samples, and achieves competitive performance against specialized tool-augmented frameworks when applied to a backbone trained only with tool-free chain-of-thought supervision.

arXiv:2607.24582v1 Announce Type: cross Abstract: Long-video understanding increasingly relies on large vision-language models and tool-augmented reasoning, but most systems apply the same inference procedure to every example regardless of difficulty. This uniform strategy invokes unnecessary tool-assisted processing for easy questions and provides limited control when difficult questions require fine-grained temporal evidence. We propose CADER Confidence-Aware Dynamic Evidence Reasoning , a training-free framework for adaptive and reliable long-video reasoning. CADER first performs global reasoning over uniformly sampled frames and estimates answer confidence with a logit-margin signal, allowing high-confidence examples to exit early. For uncertain examples, CADER activates a second-stage tool-augmented loop that combines temporal cropping, lightweight semantic verification, and Relevance-Guided Resampling to progressively localize question-relevant evidence. This design treats tool use as a sample-level decision: a single global pass handles easy cases, while additional reasoning is reserved for examples where uncertainty suggests that more evidence is needed. Experiments on multiple VideoQA benchmarks show that CADER improves long-video reasoning while bypassing Stage~2 for high-confidence samples. Moreover, when applied to a backbone trained only with tool-free chain-of-thought supervision, CADER achieves competitive performance against specialized tool-augmented frameworks, suggesting a practical inference-time route for adaptive long-video reasoning.