主动探索发现文生图模型数以十万计的未知缺陷，定位数据不足根源。
FailureAtlas:Mapping the Failure Landscape of T2I Models via Active Exploration
- 通过结构化搜索最小故障触发概念，主动挖掘模型缺陷。
- 在SD1.5中发现超24.7万条未被记录的错误片段。
- 适合研究生成模型鲁棒性与数据质量的开发者和研究员。
静态基准为文生图（T2I）模型的对比提供了重要基础，但其被动设计诊断能力有限，难以揭示系统性失败的全貌或定位根本原因。本文提出互补范式：主动探索。我们引入FailureAtlas，首个专为大规模自主探索和绘制T2I模型失败景观而设计的框架。FailureAtlas将错误发现建模为对最小故障诱导概念的结构化搜索。尽管该问题计算开销巨大，我们通过新颖加速技术使其可处理。应用于Stable Diffusion模型时，该方法发现了数十万条此前未知的错误切片（仅SD1.5即超过247,000条），并首次提供大规模证据表明这些失败与训练数据稀缺性相关。FailureAtlas作为可扩展、原则性的模型审计引擎，确立了一种以诊断为导向的新方法，推动更鲁棒生成AI的发展。代码已开源：https://github.com/cure-lab/FailureAtlas。
原文摘要 · Abstract (English)
Static benchmarks have provided a valuable foundation for comparing Text-to-Image (T2I) models. However, their passive design offers limited diagnostic power, struggling to uncover the full landscape of systematic failures or isolate their root causes. We argue for a complementary paradigm: active exploration. We introduce FailureAtlas, the first framework designed to autonomously explore and map the vast failure landscape of T2I models at scale. FailureAtlas frames error discovery as a structured search for minimal, failure-inducing concepts. While it is a computationally explosive problem, we make it tractable with novel acceleration techniques. When applied to Stable Diffusion models, our method uncovers hundreds of thousands of previously unknown error slices (over 247,000 in SD1.5 alone) and provides the first large-scale evidence linking these failures to data scarcity in the training set. By providing a principled and scalable engine for deep model auditing, FailureAtlas establishes a new, diagnostic-first methodology to guide the development of more robust generative AI. The code is available at https://github.com/cure-lab/FailureAtlas
Thank you to arXiv for use of its open access interoperability. PaperDance 不是 arXiv 官方产品；中文卡片由大模型生成，请以原文为准。