Patel Karan P, Maicke Andrew, Arzate Jared, Kwon Jaesuk, Smith J Darby, Aimone James B, Incorvia Jean Anne C, Cardwell Suma G, Schuman Catherine D
Department of Electrical Engineering and Computer Science, University of Tennessee, Knoxville, TN, USA.
Chandra Family Dept. of Electrical and Computer Engineering, The University of Texas at Austin, Austin, TX, USA.
Commun Eng. 2025 Mar 11;4(1):43. doi: 10.1038/s44172-025-00376-8.
Emerging devices, such as magnetic tunnel junctions, are key for energy-efficient, performant future computing systems. However, designing devices with the desirable specification and performance for these applications is often found to be time-consuming and non-trivial. Here, we investigate the design and optimization of spin-orbit torque and spin transfer torque magnetic tunnel junction models as the probabilistic devices for true random number generation. We leverage reinforcement learning and evolutionary optimization to vary key device and material properties of the various device models for stochastic operation. Our artificial-intelligence-guided codesign methods generated different candidate devices capable of generating stochastic samples for a desired probability distribution, while also minimizing energy usage for the devices. This framework can also be applied to other devices and applications.
诸如磁性隧道结之类的新兴器件是未来高能效、高性能计算系统的关键。然而，为这些应用设计具有理想规格和性能的器件通常既耗时又具有挑战性。在此，我们研究自旋轨道扭矩和自旋转移扭矩磁性隧道结模型的设计与优化，将其作为用于产生真随机数的概率器件。我们利用强化学习和进化优化来改变各种器件模型的关键器件和材料属性，以实现随机操作。我们的人工智能引导协同设计方法生成了不同的候选器件，这些器件能够为所需概率分布生成随机样本，同时还能将器件的能耗降至最低。该框架也可应用于其他器件和应用。
