In-memory Subnet Computation for Area and Energy Efficient AI

Shi Zhao, Yue Li, Pan Jieming, Evgeny Zamburg, Aaron Voon-Yew Thean

Keywords: Subnetwork computation, computing in-memory (CIM), Ferroelectric FET

TL;DR: We propose an in-memory computing architecture using customized dual-gate BEOL FeFETs and advanced subnet systems, enabling energy-efficient MAC operations with concurrent subnet masking for rapid, resource-efficient continual learning.

Confirmation Of Submission Requirements: I submit an abstract. It uses the template provided on the submission page and is no longer than 2 pages.

Submission Number: 142