时间如何流逝？基于心理物理学的评估用于视觉-语言模型
Which Way Does Time Flow? A Psychophysics-Grounded Evaluation for Vision-Language Models
AI总结：
本文通过心理物理学验证的基准测试评估了视觉-语言模型对时间方向的理解能力，发现大多数模型表现接近随机，且在物理不可逆过程和因果手动动作上远低于人类水平。
摘要：
现代视觉-语言模型（VLMs）在多项多模态任务中表现出色，但其对视频中时间信息的理解仍较弱且未得到充分评估。我们通过一个看似简单但具有揭示性的挑战来填补这一空白：判断时间箭头（AoT）-判断短片段是正向还是反向播放。我们引入了AoT-PsyPhyBENCH，一个心理物理学验证的基准测试，测试VLMs是否能利用与人类相同刺激和行为基线来推断自然视频中的时间方向。我们对开放权重和专有、推理和非推理VLMs的全面评估表明，大多数模型表现接近随机，即使最佳模型在物理不可逆过程（如自由落体、扩散/爆炸）和因果手动动作（除法/加法）上也远低于人类准确性。这些结果突显了当前多模态系统中的根本差距：尽管它们能捕捉丰富的视觉-语义相关性，但缺乏用于时间连续性和因果理解所需的归纳偏置。我们发布AoT-PsyPhyBENCH的代码和数据，以鼓励进一步提升VLMs在物理和时间推理能力方面的进展。
英文摘要：
Modern vision-language models (VLMs) excel at many multimodal tasks, yet their grasp of temporal information in video remains weak and has not been adequately evaluated. We probe this gap with a deceptively simple but revealing challenge: judging the arrow of time (AoT)-whether a short clip is played forward or backward. We introduce AoT-PsyPhyBENCH, a psychophysically validated benchmark that tests whether VLMs can infer temporal direction in natural videos using the same stimuli and behavioral baselines established for humans. Our comprehensive evaluation of open-weight and proprietary, reasoning and non-reasoning VLMs reveals that most models perform near chance, and even the best model lags far behind human accuracy on physically irreversible processes (e.g., free fall, diffusion/explosion) and causal manual actions (division/addition) that humans recognize almost instantly. These results highlight a fundamental gap in current multimodal systems: while they capture rich visual-semantic correlations, they lack the inductive biases required for temporal continuity and causal understanding. We release the code and data for AoT-PsyPhyBENCH to encourage further progress in the physical and temporal reasoning capabilities of VLMs.
机构、总结、中文摘要均由AI分析而来，可能存在错误，请以论文正文为准。
