Workshop on Analog Computing Opportunities

由于多种原因，模拟电路和系统近来对于主流计算变得更加重要。摩尔定律在计算中的可预见的终结就是这样一个原因，其中数字电路中过度的物理功耗所施加的限制起着重要作用。另一方面，模拟电路一直被吹捧为固有的低功耗，因此具有至少部分缓解问题的潜力。该奖项资助了一个研讨会，该研讨会将通过将领先的学者和理论家沿着行业从业者聚集在一个屋檐下，为期两天，探索模拟计算的最新技术水平。该领域的领导人将作演讲，并将举行头脑风暴会议，以确定该领域未来研究的方向。讨论将以报告的形式记录下来，这不仅将成为科学家和工程师未来工作的基础，而且还将以公共网站和国家科学基金会项目报告的形式传播。我们将邀请部分学生参加研讨会，并适当注意与会者群体的多样性，包括妇女和其他代表性不足的群体。技术的进步使大规模集成变得更加可行，模拟电路和系统也变得更加可行。虽然模拟电路固有的噪声和随之而来的不确定性仍然困扰着它们，但较新的架构和问题领域显示出可以将噪声的有害影响降至最低的希望。对模拟神经网络的更新的兴趣（使用例如，深度学习）证明了这一新兴趋势。其他技术创新，例如，用于器件和/或互连的射频技术、光学、量子和其它新兴技术也可以利用模拟技术。与数字电路相比，模拟计算的最终能力也在研究之中。虽然图灵机一直是数字计算机的理想化计算模型，并且（数字）计算机器的最终理论能力已经在这个框架中进行了评估，但类似的模拟计算研究在文献中还处于萌芽状态。例如，将计算问题的模拟复杂性分类为类似于P、NP等的类的结果是理论研究人员由于对模拟计算重新感兴趣而重新审视的一个领域。这一松散联系的研究对实践的影响--特别是在解决实际问题的硬件设计方面--仍有待观察，并将成为研讨会的核心。该奖项反映了NSF的法定使命，并被认为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来支持。

Analog circuits and systems have become more important to mainstream computing recently due to a multitude of reasons. The foreseeable end of Moore's law in computing is one such reason where limits imposed by excessive physical power dissipation in digital circuits play an important role. Analog circuits on the other hand has been touted as intrinsically low power, thus having the potential to at least partially alleviate the problem. This award funds a workshop that will explore the current state of the art in analog computing by bringing together the leading academics and theoreticians, along with industry practitioners under one roof for a period of two days. Presentations will be made by leaders in the field, and brainstorming sessions will be held to chart the direction of future research in the field. The discussions will be documented in the form of a report, which will not only form the basis for future work among scientists and engineers but will also be disseminated in the form of public websites and NSF projects reports. Selected students will be invited to participate in the workshop by paying due attention to the diversity of the group of attendees to include women and other underrepresented groups.Progress in technology enabling larger scale integration has also made analog circuits and systems more viable. While noise and attendant uncertainties inherent to the analog circuits still plague them, newer architectures and problem domains show promise where the deleterious effects of noise could be minimized. Still newer interest in analog neural networks (using e.g., deep learning) attests to this emerging trend. Other technological innovations, e.g., radio frequency techniques, optical, quantum and other emerging technologies for device and/or interconnect can also exploit analog techniques. The ultimate power of analog computing, as compared to the digital circuits is also being investigated. While the Turing machine has been a idealized model of computation for digital computers and the ultimate theoretical power of a (digital) computing machine has been assessed in this framework, similar studies for analog computing have been in nascent in the literature. For example, results classifying the analog complexity of computational problems in classes analogous to say, P, NP etc. is an area being revisited by theoretical researchers due to renewed interest in analog computing. The impact of this loosely connected body of research to practice - especially in the design of hardware for solving practical problems - remains to be seen, and will be central to the workshop.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.