Double-Gate Ferroelectric Memory for AI Acceleration
Double-Gate Ferroelectric Memory for AI Acceleration
Ferroelectric Field Effect Transistor with Multiple Independently Controlled Gates and Memory Array Device Made Thereof
Tags: National Cheng Kung University, Taiwan, Electronics & Robotics, Computing Technology
This technology involves a non-volatile semiconductor memory element using ferroelectric material as a gate-insulating layer, applied in neural network circuit arrays for AI acceleration chips. It features independently controlled double gates, allowing for data selection and storage in a single element, improving chip area efficiency. The innovation enhances traditional ferroelectric transistors, reducing the size of memory arrays while increasing energy efficiency. Applications include AI acceleration chips, particularly for neuromorphic computing and advanced memory systems. By optimizing the ferroelectric transistor structure, this technology offers improved storage density and reduced chip size.
IP Type or Form Factor: Patent Granted; Design; Material; Platform
TRL: 4 - minimum viable product built in lab
Industry or Tech Area: Semiconductors; Computing Architecture
