Alternating Perception-Reasoning for Hallucination-Resistant Video Understanding
This addresses video understanding challenges for AI systems by reducing hallucinations, though it is incremental as it builds on existing video reasoning methods.
The paper tackles the problem of hallucinations and insufficient evidence in Video Reasoning LLMs by introducing a loop-based paradigm with an anti-hallucination reward, achieving state-of-the-art results in 3B and 7B parameter scales with the best data efficiency.
Sufficient visual perception is the foundation of video reasoning. Nevertheless, existing Video Reasoning LLMs suffer from perception shortcuts, relying on a flawed single-step perception paradigm. This paradigm describes the video and then conducts reasoning, which runs the risk of insufficient evidence and emergent hallucinations. To address these issues, we introduce a new framework that integrates a loop-based paradigm with an anti-hallucination reward. First, to address the insufficient evidence, we introduce the Perception Loop Reasoning (PLR) paradigm. Instead of describing the video at once, each loop requires the model to describe a video segment with precise timestamps, analyze this segment, and decide the next action. Second, for the risk of hallucinations, the Factual-Aware Evaluator (FAE) evaluates each perception result as a reliable anti-hallucination reward. This reward encourages the model to provide sufficient and precise video evidence. Our FAE, which performs comparably to GPT-4o, is tuned on our AnetHallu-117K, a large-scale hallucination judgment preference dataset. Extensive experiments show that our Video-PLR achieves the state-of-the-art in both 3B and 7B parameter scales and has the best data efficiency. Our code, models, and datasets are released on: https://github.com/BoweiPu/VideoPLR.