AP Seminar - New Computing Paradigm based on Emerging Devices

With the development of information technology, the number of interconnected devices is rapidly increasing, accompanied by the problem of processing large amounts of unstructured data generated by these devices. Deep neural networks have shown an excellent ability to handle unstructured data. However, traditional processors based on von Neumann architecture face limitations in improving their computing speed and energy efficiency ratio when dealing with deep neural networks that have increasing numbers of parameters and operations due to the so-called von Neumann bottleneck. New computing paradigms, including computation-in-memory and in-senor computing, can achieve high-speed, low-energy consumption via in-situ computing. In this talk, I will introduce our recent research on RRAM-based CIM and FDSOI-based in-senor computing, including the device and circuit design, hardware implementation of the system, and EDA tool development.