将视觉概念合成为视觉-语言程序
Synthesizing Visual Concepts as Vision-Language Programs
把 VLM 当作结构化视觉描述生成器，再将描述编译成可在图像上直接执行的神经符号程序，为需要系统性逻辑推理的视觉任务提供了一条可解释、可抑制捷径学习的替代路径，可复用于视觉归纳推理与组合式推理基准。
原文依据 VLP leverages the model to produce structured visual descriptions that are compiled into neuro-symbolic programs.
摘要译文
由模型翻译 · 术语统一 · 请对照原文核对关键结论视觉-语言模型（VLM）在多模态任务上表现强劲，但常常在系统性视觉推理上失败，尤其是在归纳推理问题中。神经符号方法有望通过从图像中归纳出可解释的逻辑程序来解决这一问题，不过这类方法通常依赖僵化、领域特定的感知模块。我们提出视觉-语言程序（VLP），它将 VLM 的感知灵活性与符号程序合成的系统性推理结合起来。VLP 并不把推理嵌入 VLM 内部，而是利用该模型生成结构化的视觉描述，再将这些描述编译为神经符号程序。所得程序直接在图像上执行，与任务约束保持一致，并提供人类可解释的解释，从而便于缓解捷径问题。我们在合成数据集与真实数据集上的实验表明，VLP 优于对 VLM 的直接提示与结构化提示，在需要复杂逻辑推理的任务上尤为明显。
英文原文摘要 1038 字符
Vision-Language models (VLMs) achieve strong performance on multimodal tasks but often fail at systematic visual reasoning, especially in inductive reasoning problems. Neuro-symbolic methods promise to address this by inducing interpretable logical programs from images, though they usually rely on rigid, domain-specific perception modules for this. We propose Vision-Language Programs (VLP), which combine the perceptual flexibility of VLMs with the systematic reasoning of symbolic program synthesis. Rather than embedding reasoning inside the VLM, VLP leverages the model to produce structured visual descriptions that are compiled into neuro-symbolic programs. The resulting programs execute directly on images, remain consistent with task constraints, and provide human-interpretable explanations that enable easy shortcut mitigation. Our experiments across synthetic and real-world datasets demonstrate that VLPs outperform both direct and structured prompting of VLMs, particularly on tasks that require complex logical reasoning.