WiCAS
The Women in Circuits and Systems (WiCAS) Committee of the IEEE Circuits and Systems Society proudly invites you to attend the WiCAS Technical Session organized with the occasion of AICAS 2022.
WiCAS Program Schedule
The WiCAS Workshop is scheduled in June 14, 2022 (Tuesday) 12:10 (GMT +9)
Notice for AICAS 2022 registrants
- WiCAS satellite event @AICAS 2022 is available for the AICAS 2022 registrants.
- If you want to take our outstanding WiCAS lectures, please click the button below to register.
- Please be noted that all the domestic registrants are subject to attend the WiCAS sessions onsite.
WiCAS Registration
WiCAS Invited Speakers
|
Designing an Efficient HW Accelerator for AI Workloads
Dr.Soojung Ryu
CEO, SAPEON Inc., Korea
Biography Abstract
Biography
Ryu, Soojung is the Chief Executive Officer of SAPEON Inc., an US-based AI semiconductor company established through joint investment from SK Telecom, SK Square and SK Hynix. She is also serving as the CEO of SAPEON Korea. Ryu served as the Vice President and Head of AI Accelerator Office at SK Telecom. She is in charge of the development of SAPEON, SKT’s self-developed AI accelerator chip that enables the provision of high-quality AI services by enhancing the performance of AI data centers. Before joining SK Telecom in April 2021, she was a University-Industry Collaboration Professor at Seoul National University, where she conducted R&D in the neural processing unit (NPU) and processing-in-memory (PIM). From 2014 to 2018, she served as the Vice President of Samsung Electronics. She was in charge of processor design and software framework development for Samsung Reconfigurable Processor and Samsung Mobile Graphics Processing Unit. She received her Ph.D. degree in Electrical & Computer Engineering from Georgia Institute of Technology.
Abstract
The increasing demand for high-quality AI-Based services has accelerated the development of dedicated AI-specific hardware. This talk will begin with an introduction of SAPEON products in the form of chip, card, and server.
SAPEON products enables us to provide high-quality AI services to the wide public by designing its hardware architecture efficiently for AI workloads. SAPEON architecture focuses on AI inference performance by maximizing utilization of computing resources with high flexibility. High-density MAC array and programmable vector cores are orchestrated by dynamic scheduler to achieve a high performance.
In this talk, we show case studies of AI speaker, intrusion detection system, robots and more potential services in Data Centers. SAPEON’s service-oriented hardware design, software-development tools as well as AI-inference serving system will deliver the high-quality AI services in data center.
|
WiCAS 2022 Organizing Committee