Will the Era of AI Drive Emerging Technologies to Overtake CMOS?
An Chen - IBM Research, San Jose, CA
An Chen - IBM Corp., San Jose, CA
As CMOS scaling approaches its fundamental limit, numerous emerging technologies have been explored as potential beyond-CMOS solutions. Despite this extensive effort, few novel technologies have been shown to be able to surpass CMOS for Boolean logic and von Neumann architectures. The rise of Artificial Intelligence (AI) may provide new opportunities for emerging technologies, with significantly different applications, algorithms, architectures, and device requirements. Although customized CMOS chips can perform better than general-purpose processors for AI tasks, they are still orders of magnitude poorer than biologic brains in terms of efficiency and cognitive capabilities. While some emerging devices have shown promising characteristics to mimic neuron and synapsis, there are obvious shortcoming of these devices for large scale neural networks. This panel will bring industry and academic experts to address this question: Will the era of AI provide an opportunity for emerging technologies to overtake CMOS?
Kaushik Roy - Purdue Univ., West Lafayette, IN Meng-Fan Chang - National Tsing Hua Univ., Hsinchu City, Taiwan Hsien-Hsin Sean Lee - Taiwan Semiconductor Manufacturing Co., Ltd., Hsinchu, Taiwan Geoffrey W. Burr - IBM Research - Almaden, San Jose, CA