神经形态工程学
横杆开关
材料科学
纳米技术
光电子学
计算机体系结构
计算机科学
电信
人工智能
人工神经网络
作者
Setareh Kazemzadeh,Tim Stevens,Yoeri van de Burgt
标识
DOI:10.1002/aelm.202500054
摘要
Abstract Analog non‐volatile memory devices, such as electrochemical random‐access memory (ECRAM), have emerged as a promising platform for in‐memory computing, facilitating efficient data processing. In this research, a pioneering approach is presented by introducing an all‐organic and fully integrated crossbar array comprising 3 × 3 ECRAM devices, notable for its facile fabrication employing photolithography techniques and exclusive utilization of organic materials for the devices and resistors. The crossbar array demonstrates remarkable capabilities, enabling inference and in situ parallel training, leading to high accuracy when classifying linearly separable 2D and 3D tasks. Notably, the biocompatible nature of the materials employed in the array offers promising prospects for the development of smart and adaptable bioelectronics that can directly interface with the biological environment.
科研通智能强力驱动
Strongly Powered by AbleSci AI