arXiv 2506.220752026-07-01cs.CV

通过快速与慢速思考学习进行机器视觉推理

Reasoning in machine vision by learning fast and slow thinking

Centre for Bioengineering, Queen Mary University of London（伦敦玛丽女王大学生物工程中心）；Digital Environment Research Institute, Queen Mary University of London（伦敦玛丽女王大学数字环境研究所）；School of Engineering and Materials Science, Queen Mary University of London（伦敦玛丽女王大学工程与材料科学学院）；UCL Hawkes Institute, University College London（伦敦大学学院UCL霍克斯研究所）；Department of Medical Physics and Biomedical Engineering, University College London（伦敦大学学院医学物理与生物医学工程系）；Centre for Urology Imaging, Prostate, AI and Surgical Studies (COMPASS) Research Group, Division of Surgery and Interventional Science, University College London（伦敦大学学院外科与介入科学部泌尿影像、前列腺、人工智能与手术研究（COMPASS）组中心）；Department of Urology, Comprehensive Cancer Center, Medical University of Vienna（维也纳医科大学综合癌症中心泌尿科）；Division of Surgery and Interventional Science, University College London（伦敦大学学院外科与介入科学部）；Department of Computer Science, University College London（伦敦大学学院计算机科学系）

Shaheer U. Saeed, Yipei Wang, Veeru Kasivisvanathan, Brian R. Davidson, Matthew J. Clarkson, Yipeng Hu, Daniel C. Alexander

发表信息：: Nat Commun (2026)
DOI：: 10.1038/s41467-026-74579-8

阅读 PDF

arXiv 页面 ↗HTML 原文 ↗

AI总结：

受人类双过程认知理论启发，提出结合快速系统I和慢速系统II的视觉推理范式，通过推理时计算提升数据稀缺场景下的性能，在计算机视觉基准和五器官癌症定位任务中超越大规模监督学习和人类专家。

摘要：

推理是人类智能的标志，能够在复杂的不熟悉场景中实现自适应决策。相比之下，机器智能仍然受限于训练数据，无法在推理时动态优化解决方案。尽管最近的进展探索了机器推理——用推理时计算换取性能提升——但它们专注于语言领域，如数学问题求解，其中显式规则指导逐步解决方案的生成。许多任务缺乏足够的标注数据，需要替代的性能提升机制，例如推理时计算。在这里，我们提出了一种视觉机器推理范式，即使在标注数据有限的情况下，也能随着思考时间（推理时计算）的增加而实现性能提升。我们的方法受人类认知双过程理论的启发，集成了一个快速思考的系统I模块，用于在熟悉任务中生成和验证解决方案，以及一个慢速思考的系统II模块，通过自我对弈强化学习迭代优化预测，即使任务特定数据有限。该范式涉及提出、竞争和优化解决方案，直到收敛。我们证明，与大规模监督学习、基础模型和人类专家在视觉任务中相比，扩展的推理时计算能产生更优的性能。这些任务包括计算机视觉基准和跨五个器官的癌症定位，突显了推理时计算在数据稀缺问题中的潜力。

英文摘要：

Reasoning is a hallmark of human intelligence, enabling adaptive decision-making in complex unfamiliar scenarios. In contrast, machine intelligence remains bound to training data, unable to dynamically refine solutions at inference. While recent advances have explored machine reasoning - trading inference-time compute for improved performance - they focus on verbal domains such as mathematical problem-solving where explicit rules govern step-by-step solution generation. Many tasks lack sufficient labelled data and require alternative performance improvement mechanisms, such as inference-time compute. Here we present a paradigm for machine reasoning in vision, enabling performance improvements with increasing thinking time (inference-time compute), even with limited labelled data. Our approach is inspired by dual-process theories of human cognition, integrating a fast-thinking System I module for generating and verifying solutions in familiar tasks, with a slow-thinking System II module that iteratively refines predictions using self-play reinforcement learning, even when task-specific data is limited. This paradigm involves proposing, competing over, and refining solutions until convergence. We demonstrate that extended inference-time compute yields superior performance compared to large-scale supervised learning, foundation models, and human experts in vision tasks. These include computer-vision benchmarks and cancer localisation across five organs, highlighting the potential of inference-time compute for data-scarce problems.

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。