【直播】【大湾区自旋量子科技论坛】Prof. Johan Åkerman 学术报告

直播信息

报告题目

Spin Hall nano-oscillator based Ising Machines for combinatorial optimization

报告人

Prof. Johan Åkerman

报告时间

2022年6月22日 20:00

主办方

香港中文大学(深圳)

深圳技术大学

华南师范大学

松山湖材料实验室

直播二维码

直播海报

报告人介绍

Prof. Johan Åkerman received his Ph.D. in Materials Physics from KTH Royal Institute of Technology in 2000. After a post-doc at University of California, San Diego, he joined Motorola, for four years, where he was responsible for MRAM reliability. The MRAM technology he helped to launch remains the most commercially successfully MRAM to date. In 2005, he returned to Sweden to start his own research group at the Department of Materials and Nanophysics at KTH Royal Institute of Technology. In 2008 he was recruited as Full Professor to the Physics Department at University of Gothenburg, while remaining a Guest Professor at KTH. Prof. Åkerman has been working with spintronic technology for the last 20 years and has authored over 250 scientific papers with over 13 000 citations. He holds close to 20 patents and is the founder of two start-up companies, NanOsc AB, commercializing spintronic devices, and NanOsc Instruments AB, designing and manufacturing spectrometers for ferromagnetic resonance measurements at cryogenic and room temperatures. His main projects are related to spin torque and spin Hall nano-oscillators, with particular focus on mutual synchronization, magnetodynamical solitons, and oscillator networks for neuromorphic computing.

报告摘要

Figure 1 

a) A voltage-gated SHNO where the E-field controls the PMA of the CoFeB layer. 

b) A memristor-controlled SHNO where an additional memristor current adds or subtracts to the SHNO drive current.

In my talk I will first briefly review the fundamentals of Ising Machines and describe different types of quantum and classical implementations. These range from D-waves Quantum Annealer, so-called Coherent Ising Machines based on Degenerate Optical Parametric Oscillators (DOPO), and most recently breadboard and PCB based LCR oscillator networks. I will then describe our recent progress in mutually synchronized nano-constriction based spin Hall nano-oscillator (SHNO) chains and twodimensional arrays and the possibility of using them for Ising Machines. I will describe how we can generate different types of spin wave modes and control those with voltage cates. In our most recent work, we have now added memristive functionality to these SHNOs and demonstrated non-volatile control of mutual synchronization and how this can be used for neuromorphic computing. Using microwave current injection locking and phase-sensitive micro-Brillouin Light Scattering microscopy, such two-dimensional SHNO arrays can be efficiently phasebinarized, which is a prerequisite for oscillator based Ising Machines, and solve MAXCUT problems.

扩展阅读

 

编辑：黄琦

蔻享学术平台，国内领先的一站式科学资源共享平台，依托国内外一流科研院所、高等院校和企业的科研力量，聚焦前沿科学，以优化科研创新环境、传播和服务科学、促进学科交叉融合为宗旨，打造优质学术资源的共享数据平台。

版权说明：未经授权严禁任何形式的媒体转载和摘编，并且严禁转载至微信以外的平台！

---

原创文章首发于蔻享学术，仅代表作者观点，不代表蔻享学术立场。

转载请在『蔻享学术』公众号后台留言。

点击阅读原文～发现惊喜！