Sapienz Receives ISSTA 2026 Impact Paper Award

Ke Mao, Mark Harman, and Yue Jia have been recognized with the ACM SIGSOFT International Symposium on Software Testing and Analysis (ISSTA) 2026 Most Influential Paper Award for their 2016 research, “Sapienz: multi-objective automated testing for Android applications.” The recognition celebrates a decade of impact from work that began at University College London and went on to become testing infrastructure at Meta.

Ke Mao、Mark Harman 和 Yue Jia 凭借其 2016 年的研究成果“Sapienz：面向 Android 应用程序的多目标自动化测试”荣获 ACM SIGSOFT 软件测试与分析国际研讨会 (ISSTA) 2026 年度最具影响力论文奖。该奖项旨在表彰这项始于伦敦大学学院、随后成为 Meta 测试基础设施的研究在过去十年中所产生的影响。

Sapienz automatically tests apps to find crashes and helps engineers reproduce them in fewer steps, making failures easier to investigate. The original research combined three goals: exploring more of an app, finding more crashes, and reducing the number of steps needed to trigger them.

Sapienz 通过自动测试应用程序来发现崩溃，并帮助工程师以更少的步骤重现这些崩溃，从而使故障调查变得更加容易。最初的研究结合了三个目标：探索更多的应用程序功能、发现更多的崩溃，以及减少触发这些崩溃所需的步骤。

In its evaluation on 1,000 popular Google Play apps, Sapienz found 558 previously unknown crashes. Within 17 months of publication, Sapienz was deployed at scale at Meta. Today, it runs millions of tests daily across Meta’s platforms, surfaces thousands of unique crashes each week, and helps safeguard hundreds of app releases each month, supporting product reliability and quality.

在对 1,000 款热门 Google Play 应用程序的评估中，Sapienz 发现了 558 个此前未知的崩溃。在论文发表后的 17 个月内，Sapienz 便已在 Meta 实现了大规模部署。如今，它每天在 Meta 的各个平台上运行数百万次测试，每周发现数千个独特的崩溃，并每月协助保障数百个应用程序版本的发布，从而支持产品的可靠性和质量。

Sapienz’s applications quickly expanded beyond crash detection. By using test accounts with synthetic conversations, it became the app-exploration engine for dynamic taint analysis at WhatsApp. We shared Sapienz’s journey in October in our ISSTA 2026 keynote delivered by Ke Mao, “Finding the Bugs Users Would Find: From Sapienz to Autonomous Agents.”

Sapienz 的应用很快扩展到了崩溃检测之外。通过使用带有合成对话的测试账户，它成为了 WhatsApp 动态污点分析的应用程序探索引擎。我们在 10 月份由 Ke Mao 发表的 ISSTA 2026 主题演讲“寻找用户会发现的漏洞：从 Sapienz 到自主智能体”中分享了 Sapienz 的历程。

Beyond detecting crashes, large language model (LLM) agents can now explore apps and reason about whether their behavior matches what users should expect, extending automated testing from reliability to functional correctness and, increasingly, automated repair. Sapienz’s core ideas of multi-objective exploration and actionable fault revelation carry over directly. A decade on, those founding principles still hold true.

除了检测崩溃之外，大语言模型 (LLM) 智能体现在可以探索应用程序，并推断其行为是否符合用户的预期，从而将自动化测试从可靠性扩展到功能正确性，并日益扩展到自动化修复。Sapienz 关于多目标探索和可操作故障揭示的核心理念直接延续了下来。十年过去了，这些基本原则依然适用。

As part of Meta’s overall privacy and risk management program, this work reflects our ongoing efforts to ensure privacy considerations are at the heart of our product development process, helping teams identify risks earlier in the development lifecycle, apply safeguards more consistently, and build privacy protections into each new innovation from the start.

作为 Meta 整体隐私和风险管理计划的一部分，这项工作反映了我们为确保隐私考量处于产品开发流程核心地位所做的持续努力，帮助团队在开发生命周期的早期识别风险，更一致地应用保障措施，并将隐私保护从一开始就融入到每一项新的创新中。