基于相变忆阻器的亚10毫秒神经动力学系统

Lei Cai1†, Yaoyu Tao1,2,3*†, Chenchen Xie4†, Longhao Yan1†, Shiqian Li2, Ruihong Shen2, Zelun Pan1, Xile Wang1, Bowen Wang1, Daijing Shi1, Yihang Zhu1, Teng Zhang1, Yixin Zhu2,3*, Xi Li4, Zhitang Song4*, Ru Huang1,2, Yuchao Yang1,2,5,6* High-fidelity geometry for physical-world modeling demands real-time, dense, and differentiable deformation fields on manifolds. Neural dynamical systems (NDSs) using adaptive stepsize integration with embedded neural networks excel at these tasks but still suffer latency on the order of hundreds of milliseconds. In this work, we report a sub–10-millisecond NDS hardware leveraging the precisely controlled conductance drift of phase-change memristors and their multilevel compute-in-memory capabilities. We fabricated a 40-nanometer NDS chip for the challenging surface reconstruction tasks. …