2019年10月24日，秋高气爽，丹桂飘香，电子科学与应用物理学院有幸请到工作在大洋彼岸的美国圣母大学的胡小波教授，为我院师生做“In-Memory Computing for Machine Learning Applications and Beyond”的学术报告。
报告伊始，电物学院院长——梁华国教授为在座师生介绍胡小波教授的研究领域与个人经历。胡教授是美国圣母大学计算机科学与工程系教授。她的研究领域包括低功耗系统设计、新兴技术的电路和架构设计、实时嵌入式系统和软硬件协同设计。她在这些领域发表了300多篇论文，获得过设计自动化会议和低功率电子与设计国际研讨会的最佳论文奖，以及NSF职业奖。胡教授参与了几个大型企业和政府资助的中心级项目，是NSF/SRC E2CDA项目的主题负责人，也是设计自动化会议的总主席和项目主席，还是2019年IEEE实时系统研讨会的项目主席。她还担任过IEEE关于VLSI的会刊、ACM关于电子系统设计自动化的会刊等的副编辑，还是ACM关于网络物理系统的会刊的副编辑。
紧接着，胡教授通过引入传统计算存储器的相关背景，提出了一种全新的硬件设计——FETs。本次报告介绍了一种基于铁电场效应晶体管(一种新兴的非易失性器件)的内存计算模块的跨层设计。FeFET是将铁电材料层集成在MOSFET的栅栈中而制成的，它既可以作为晶体管，又可以作为非易失性存储元件。这一独特的性能使的器件具有高效局域性和低功耗精细集成的逻辑和内存。胡教授阐述了基于FeFETs的新型电路架构，以实现内存计算和可寻址内存。在应用方面在多个方向上较其他技术上具有较高的可实现性和优良性能，特别是在机器学习这一人工智能领域优点突出。内容充实同时幽默风趣激起了广大师生的兴趣，获得在座师生的高度赞扬。
在现场提问环节，与会的各位教授、老师对报告内容产生浓厚的兴趣，提出了许多与各自领域相关的学术问题，与胡教授交换了各自的意见与建议。大家围绕FeFET的安全性，可靠性进行了深入讨论，拓宽下一个研究点，该报告内容充实，研究新颖，逻辑严谨，理论扎实，在座师生受益匪浅。非常感谢胡小波教授的学术汇报，再会。
报告人简介
X. Sharon Hu is a professor in the department of Computer Science and Engineering at the University of Notre Dame, USA. Her research interests include low-power system design, circuit and architecture design with emerging technologies, real-time embedded systems and hardware-software co-design. She has published more than 300 papers in these areas. Some of her recognitions include the Best Paper Award from the Design Automation Conference and from the International Symposium on Low Power Electronics and Design, and the NSF CAREER award. She has participated in several large industry and government sponsored center-level projects and is a theme leader in an NSF/SRC E2CDA project. She served as the General Chair and Program Chair of Design Automation Conference and is the Program Chair of 2019 IEEE Real-Time Systems Symposium. She also served as Associate Editor for IEEE Transactions on VLSI, ACM Transactions on Design Automation of Electronic Systems, etc. and is an Associate Editor of ACM Transactions on Cyber-Physical Systems. X. Sharon Hu is a Fellow of the IEEE.
报告简介
Data transfer between a processor and memory is a major bottleneck in improving application-level performance. This is particularly the case for data intensive tasks such as some machine learning applications. In-memory computing, where certain data processing is performed in memory, could be an effective solution to address this bottleneck. Consequently, compact, low-power, fast and non-volatile in-memory computing is highly desirable. This talk presents a cross-layer effort of designing in-memory computing modules based on ferroelectric FETs, an emerging, non-volatile device. An FeFET is made by integrating a ferroelectric material layer in the gate stack of a MOSFET, and can behave as both a transistor and a non-volatile storage element. This unique property enables area efficient and low-power finely integrated logic and memory. Novel circuits/architectures based on FeFETs to accomplish computing in memory, content addressable memory and crossbar arrays will be elaborated. Application-level benefits, particularly for machine learning, in comparison with other alternative technologies will be discussed.
（文/解为强）
