Next CRC-colloquium on August 22, 2024: Duygu Kuzum (Jacobs School of Engineering, University of California, San Diego, USA)
Engineering Nanodevices for Energy Efficient Analog
Compute In Memory
Prof. Dr. Duygu Kuzum
Jacobs School of Engineering, University of California,
San Diego, USA
Emerging nanoelectronic devices offer an attractive solution to minimize off-chip data transfer and parallelize on-chip computations for modern data-intensive AI models. In this talk, I will present our work on forming-free bulk RRAM arrays, and Mott activation neurons for scalable, large-scale implementation of analog compute in memory. I will discuss how we optimize each technology and follow a hardware/software co-design approach for integration to improve computational efficiency. I will conclude my talk with future directions and applications of neuromorphic computing in neural sensing with an emphasis on novel electronic materials and computing paradigms based on machine learning.
CRC 1461: Neurotronics
Colloquium: 22-August-2024_30, Thursday, 3:30 pm to 4:00 pm (CET)
The colloquium will start at 4:00 pm
Invited by Sonja Reich, referral by Hermann Kohlstedt
Kiel University, Faculty of Engineering