CAREER: iMPACT: Metaphysical and Probabilistic-Based Computing Transformation with Emerging Spin-Transfer Torque Device Technology

有效应对即将到来的“zettabyte”数据爆炸需要计算能力和能源效率的巨大飞跃。然而，随着摩尔定律的迅速缩小，如果传统的“确定性和精确性”计算模型仍然占主导地位，CMOS技术的物理限制使得实现高能效几乎不可能。这个CAREER提案的灵感来自于类脑计算的应用，旨在开发一种替代的、非布尔的、非CMOS计算范式，能够克服量子相关器件物理学对数字CMOS技术计算效率的限制。这项研究如果成功，将为应对即将到来的“zettabyte”数据爆炸提供解决方案。此外，PI以自下而上（公共STEM教育）和自上而下（博士生招聘）的方式解决扩大代表性不足的少数群体参与的挑战。特别是，PI和他的团队将利用奥兰多科学中心作为主要平台，通过创新的展览和小型讲座激发公众的兴趣，并在计算的背景下参与STEM教育。这个CAREER提案围绕着形而上学和概率计算转换，这是一种创新的范式，通过利用极端条件下的CMOS器件或新兴的自旋扭矩器件中的随机驱动物理现象，超越了确定性计算。最终目标是实现，对于广泛的基于感知的计算应用程序-因此对于人类大脑比当前机器表现更好的应用程序-计算性能提高10倍以上，能源效率提高100倍以上，硬件鲁棒性提高10倍以上。

Effectively tackling the upcoming "zettabyte" data explosion requires a quantum leap in computing power and energy efficiency. However, with the Moore's law dwindling quickly, the physical limits of CMOS technology make it almost impossible to achieve high energy efficiency if the traditional "deterministic and precise" computing model still dominates. This CAREER proposal is inspired by applications in brain-like computing and aims at developing an alternative, non-Boolean, non-CMOS computing paradigm capable of overcoming the limitations in computational efficiency of digital CMOS technology posed by quantum-related device physics. This research, if successful, will offer a solution to tackle the upcoming "zettabyte" data explosion. In addition, the PI approaches the challenge of broadening participation from underrepresented minority groups in both bottom-up (public STEM education) and top-down (PhD students recruiting) manner. In particular, the PI and his team will use Orlando Science Center as the main platform to stimulate public interests through innovative exhibits and mini-lectures as well as engage in STEM education in the context of computing. This CAREER proposal revolves around the Metaphysical and Probabilistic Computing Transformation, an innovative paradigm that transcends deterministic computing by natively exploiting randomness-driven physical phenomenon, either from CMOS devices under extreme conditions or from emerging spin-torque devices. The ultimate goal is to achieve, for a wide-range of perception-based computing applications - and thus for applications where the human Brain performs better than current machines - more than 10 times improvement in computing performance, more than 100 times in energy efficiency, and more than 10 times in hardware robustness over the existing state-of-the-art.