迈向以多模态大语言模型为副驾驶的可扩展网络无障碍审计
Towards Scalable Web Accessibility Audit with MLLMs as Copilots
AI总结：
本文提出AAA审计框架，通过GRASP图采样与MaC多模态大语言模型副驾驶实现WCAG-EM的可扩展人机协作执行，并贡献四个基准数据集验证方法有效性。
摘要：
确保网络无障碍对于在数字空间中推进社会福利、正义与平等至关重要，然而绝大多数网站用户界面仍然不符合规范，部分原因在于当前审计实践资源密集且难以扩展。虽然WCAG-EM为站点级一致性评估提供了一套结构化方法，但它需要大量人力投入，并且缺乏在大规模执行方面的实际支持。在这项工作中，我们提出了一个审计框架AAA，通过人机协作模式将WCAG-EM付诸实施。AAA以两项关键创新为支撑：GRASP，一种基于图的多模态采样方法，通过学习视觉、文本和关系线索的嵌入来确保具有代表性的页面覆盖；以及MaC，一种基于多模态大语言模型的副驾驶，通过跨模态推理和对高投入任务的智能辅助来支持审计人员。这些组件共同实现了可扩展的端到端网络无障碍审计，使人类审计人员能够借助AI增强的辅助产生现实世界的影响。我们还贡献了四个新颖的数据集，用于对审计流程的核心阶段进行基准测试。大量实验证明了我们方法的有效性，并提供了见解：小规模语言模型在微调后可以充当有能力的专家。
英文摘要：
Ensuring web accessibility is crucial for advancing social welfare, justice, and equality in digital spaces, yet the vast majority of website user interfaces remain non-compliant, due in part to the resource-intensive and unscalable nature of current auditing practices. While WCAG-EM offers a structured methodology for site-wise conformance evaluation, it involves great human efforts and lacks practical support for execution at scale. In this work, we present an auditing framework, AAA, which operationalizes WCAG-EM through a human-AI partnership model. AAA is anchored by two key innovations: GRASP, a graph-based multimodal sampling method that ensures representative page coverage via learned embeddings of visual, textual, and relational cues; and MaC, a multimodal large language model-based copilot that supports auditors through cross-modal reasoning and intelligent assistance in high-effort tasks. Together, these components enable scalable, end-to-end web accessibility auditing, empowering human auditors with AI-enhanced assistance for real-world impact. We further contribute four novel datasets designed for benchmarking core stages of the audit pipeline. Extensive experiments demonstrate the effectiveness of our methods, providing insights that small-scale language models can serve as capable experts when fine-tuned.
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。
