NeTS: Small: Collaborative Research: Towards Adaptive and Efficient Wireless Computing Networks

今天的移动设备不仅是智能的，而且随着Facebook Caffe2和谷歌张量流Lite等人工智能应用被推向移动设备，以及移动设备正被集成到云-雾-移动架构中，它们正在变得智能。这要求无线网络的高效和自适应的计算/通信联合设计，以优化应用级延迟(包括通信延迟和计算时间)，并实现能源效率(考虑通信和计算的能量消耗)。该项目开发低延迟、高能效和以计算为中心的无线网络的基础理论和新架构，以支持新兴的移动智能应用。私人投资机构开发的理论和算法不断融入两所大学教授的本科生和研究生课程。该项目还为本科生和高中生提供了使用最先进的无线技术的实践体验。计算/通信联合设计虽然对无线网络来说是新事物，但在数据中心网络中却是一个中心话题。然而，由于无线网络的独特特性，如无线干扰、信道衰落和有限的能量，所提出的解决方案虽然鼓舞人心，但并不直接适用于无线计算网络。该项目专注于动态和自适应地调度计算任务和数据传输以满足应用级性能要求的可证明最优机制，并由三个相互依赖的推力组成：(I)最优计算/通信联合设计。这一推力开发了无线计算网络的数学模型和理论限制，(Ii)稳健的计算/通信联合设计。这一重点集中于健壮的计算和通信联合设计，在不完美的状态信息和不可避免的短期系统过载情况下获得期望的性能，以及(Iii)学习辅助的自适应计算/通信联合设计。这一努力通过利用历史数据和可预测的用户行为进一步提高了无线计算网络的性能。这一奖项反映了NSF的法定使命，并通过使用基金会的智力优势和更广泛的影响审查标准进行评估，被认为值得支持。

Today's mobile devices are not merely smart, they are becoming intelligent as artificial intelligence applications such as Facebook Caffe2 and Google Tensor-Flow Lite are being pushed into mobile devices and as mobiles devices are being integrated into the cloud-fog-mobile architecture. This calls for efficient and adaptive computing/communication co-design of wireless networks to optimize application-level latency (including both communication latency and computing times) and to achieve energy efficiency (considering energy consumed by both communications and computing). This project develops fundamental theories and novel architectures of low-latency, energy-efficient, and computing-centric wireless networks to support emerging mobile intelligence applications. Theories and algorithms developed by the PIs are constantly integrated into the undergraduate and graduate courses taught at the two universities. This project also provides hand-on experiences to undergraduate and high school students with state-of-the-art wireless technologies. Computing/communication co-design, while new for wireless networks, is a central topic in data center networks. However, the proposed solutions, while inspiring, are not directly applicable to wireless computing networks because of the unique features of wireless networks such as wireless interference, channel fading and limited energy. This project focuses on provably optimal mechanisms that dynamically and adaptively schedule computing tasks and data transmissions to meet application-level performance requirements, and consists of three interdependent thrusts: (i) Optimal computing/communication co-design. This thrust develops mathematical models and theoretical limits of wireless computing networks, (ii) Robust computing/communication co-design. This thrust focuses on robust computing and communication co-design that achieves desired performance with imperfect state information and under unavoidable short-term system overload and (iii) Learning-aided adaptive computing/communication co-design. This thrust further improves the performance of wireless computing networks by leveraging both historical data and predictable user behaviors.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.