Verification, Validation, and Test of ML Systems (V-TML) Workshop

电子系统正在许多行业创造新的创新周期，包括汽车和制造业。这些系统是自治系统等领域创新的核心，通常包含机器学习（ML）组件来做出决策。由于安全性、可靠性和可预测性在这些系统中至关重要，该奖项资助了一个研讨会，以探索机器学习方法和系统的安全、可靠和可预测的应用，其性能通常取决于用户或系统设计师无法完全控制的因素。这些因素可能包括训练数据的不确定性，缺乏强大的算法，以及推理机制的概率方面等。组织委员会以及参与者由学术界，工业界和政府的个人组成，包括来自少数民族，妇女和代表性不足的群体的各种参与。研讨会上讨论的技术方面的例子包括但不限于：（1）相关机器学习技术的调查，（2）电子设计自动化目前使用的数据分析，以及使用类似范例的问题，以及（3）ML背景下的鲁棒性，测试，验证和验证。研讨会的成果将被公开报道，并与研究界广泛分享。该奖项反映了NSF的法定使命，并被认为是值得通过使用基金会的知识价值和更广泛的影响审查标准进行评估的支持。

Electronic systems are creating a renewed innovation cycle in many industries, including automobiles and manufacturing. These systems, which are at the heart of innovation in areas such as autonomous systems, often incorporate Machine Learning (ML) components to make decisions. Since safety, reliability, and predictability are paramount in these systems, this award funds a workshop to explore safe, reliable, and predictable applications of machine learning methods and systems, the performance of which often depends on factors that are not within the full control of the user or the system designer. These latter factors could range from uncertainties in training data, to lack of robust algorithms, to probabilistic aspects of inference mechanism etc. The organizing committee, as well as the participants are composed of individuals from academia, industry and the government, include diverse participation from minority, women and underrepresented groups.Examples of technical aspects to be discussed at the workshop include, but are not limited to: (1) survey of relevant machine learning techniques, (2) data analytics currently used by electronic design automation, and problems in which similar paradigms are used, and (3) robustness, test, validation and verification in the context of ML. The outcomes of the workshop will be reported publicly and shared widely with the research community.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.