Paper Title: Analog Computing for AI/ML Using Embedded Flash

Paper Abstract: Differently from the high density embedded Non-Volatile Memory (eNVM) technologies such as dual-poly embedded Flash (eFlash), FeRAM, STT-MRAM, and RRAM that typically require process overhead beyond standard logic process, single-poly eFlash can be built in a standard logic process without process overhead. The first part of this talk reviews various single-poly eFlash options and introduce our differentiated single-poly eFlash technology implemented in a generic logic process for moderate density eNVM applications. Recently, multi-level-cell analog flash memory has gained a new momentum as a non-volatile analog memory for the emerging artificial intelligence and security applications. For successful deployment in these new applications, however, flash memory cell needs to be programmed precisely at the target analog level. The second part of this talk, we review several previous analog flash memory approaches and briefly introduce our silicon proven fine grained analog programming scheme for flash memory based analog computing applications.

Paper Author: Seung-hwan Song, CTO and Co-Founder, ANAFLASH

Author Bio: Dr. Seung-hwan Song is a CTO and Co-Founder of Anaflash (www.anaflash.com), that develops embedded flash memory based AI solution since 2017. Previously, he has held various research and development positions at HGST (acquired by WD), Seagate, Qualcomm, Broadcom, and Samsung. He received the B.S. and M.S. degrees in electrical engineering from Seoul National University, in 2004 and 2006, and the Ph. D. degree in electrical engineering (with a minor in biomedical engineering) from University of Minnesota, in 2013. Dr Song has 45+ US patents granted around SSD, NVM, and Flash memory, and has received a best paper award at IEEE ICC 2016 and a low power design contest award at IEEE ISLPED 2012.