CAREER:Novel Nanoelectronic Reconfigurable Synaptic Memory Devices

职业：新型纳米电子可重构突触存储设备

• 批准号: 1556294
• 负责人: Rashmi Jha
• 金额: $ 9.2万
• 依托单位: University of Cincinnati Main Campus
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2015
• 资助国家: 美国
• 起止时间: 2015-08-15 至 2019-12-31
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=1556294&HistoricalAwards=false
• 关键词: CAREER; Novel; Nanoelectronic; Reconfigurable; Synaptic

Intellectual Merit: The ability to perform low-power, extreme-scale, robust neuromorphic computing, inspired by a biological brain, has eluded the scientific community for a long time. A major roadblock lies in achieving the stringent power vs. area requirement for nanoelectronic devices to implement biologically inspired computational paradigms manifested by biological synapses on a silicon chip. The proposed research aims to develop ultra-low power synaptic memory devices based on doped transition metal oxide that can mimic the transient and steady-state dynamics of a biological synapse. This objective will be accomplished by executing the specific aims including: (1) development of synaptic memory devices based on doped HfO2 stacks, (2) development of bidirectional selector diode, and (3) study of integrated synaptic crossbar networks with Complementary Metal Oxide Semiconductor (CMOS) neurons as well as biological neurons. This research will present a platform device technology to achieve dynamic synaptic response in two-terminal devices fabricated using HfO2 by selective defect engineering through controlled doping, provide a fundamental understanding of the characteristic of defects responsible for synaptic behavior and device models, and extend the application of these device for synaptic computation and learning in neural circuits, neurophysiology, and finally the neuroengineering. The successful completion of this project will have transformative impact towards enabling low-power extreme-scale computing inspired by the massive parallelism of a biological-brain that has been identified as a scientific grand challenge by the US Department of Energy and National Academy of Engineering.Broader Impacts: This project will strengthen collaborations with local small businesses and Semiconductor companies, which will have a significant impact on graduate and undergraduate training, workforce development, transition of research into advanced technology product, and regional economic growth. On educational front, the proposed research will train graduate and undergraduate students in the areas of micro/nano-electronic device fabrication, electrical testing, and data analysis and involve high-school girls in research through summer internship programs. The research findings will be tightly integrated in existing courses at graduate and undergraduate levels, and in the new course titled 'NanoScience for Electronic Era' that will be developed. As a part of this course, a teaching module on memristive devices will be developed with hands-on experiments that will be implemented in relevant courses at the University of Toledo and at Ohio Northern University which is primarily an undergraduate program school, and also made available openly for use at the other universities. The participation of women in engineering will be encouraged through programs like 'Women in Science Day of Meeting (WISDOM)', and participation of other underrepresented groups in engineering will be encouraged by interacting with Toledo Public Schools and programs like 'Engineers for a Day' and 'Excel' at the University of Toledo. To reach out to the underrepresented groups of students (mostly Hispanic) in Ohio's rural communities, the PI will continue her involvement in the course titled 'Using Sustainable Engineering Activities to Teach Science', a distance-learning course that provides professional development and training to the teachers in rural Ohio who then teach mathematics and science to the children of migrant farm workers in Ohio. The results will be promptly disseminated to the scientific community through publications in peer-reviewed journals and conference presentations, and presented to general community through educational TV channel Plugged-In and YouTube videos.

智力优势：在生物大脑的启发下，执行低功耗、极端规模、强大的神经形态计算的能力，长期以来一直困扰着科学界。一个主要的障碍在于实现纳米电子器件严格的功率与面积要求，以实现由硅芯片上的生物突触所体现的生物启发的计算范式。该研究旨在开发基于掺杂过渡金属氧化物的超低功耗突触记忆器件，该器件可以模拟生物突触的瞬态和稳态动力学。这一目标将通过执行具体目标来实现，包括：(1)开发基于掺杂HfO2堆栈的突触记忆器件，(2)开发双向选择二极管，以及(3)研究具有互补金属氧化物半导体（CMOS）神经元和生物神经元的集成突触交叉杆网络。本研究将提出一种通过控制掺杂的选择性缺陷工程技术，在HfO2制备的双端器件中实现动态突触响应的平台器件技术，为突触行为缺陷的特征和器件模型提供基本的认识，并扩展这些器件在神经回路、神经生理学、最终神经工程中突触计算和学习的应用。该项目的成功完成将对实现低功耗极端规模计算产生变革性影响，这种计算受到生物大脑的大规模并行性的启发，这已被美国能源部和国家工程院确定为科学上的重大挑战。更广泛的影响：该项目将加强与当地小企业和半导体公司的合作，这将对研究生和本科生培训、劳动力发展、研究向先进技术产品的转变以及区域经济增长产生重大影响。在教育方面，拟议的研究将培养微/纳米电子器件制造、电气测试和数据分析领域的研究生和本科生，并通过暑期实习项目让高中女生参与研究。研究成果将紧密地整合到研究生和本科生的现有课程中，以及将开发的名为“电子时代的纳米科学”的新课程中。作为本课程的一部分，将开发一个关于记忆装置的教学模块，并进行实践实验，该模块将在托莱多大学和俄亥俄北方大学（主要是本科项目学校）的相关课程中实施，并开放给其他大学使用。通过“女性科学日会议”（women in Science Day of Meeting，简称WISDOM）等项目鼓励女性参与工程领域，通过与托莱多公立学校以及托莱多大学的“工程师日”和“Excel”等项目的互动，鼓励其他代表性不足的群体参与工程领域。为了接触到俄亥俄州农村地区未被充分代表的学生群体（主要是西班牙裔），PI将继续参与名为“使用可持续工程活动来教授科学”的课程，这是一门远程学习课程，为俄亥俄州农村地区的教师提供专业发展和培训，然后他们向俄亥俄州农场移民工人的孩子教授数学和科学。研究结果将通过同行评议的期刊和会议报告迅速传播给科学界，并通过教育电视频道plug - in和YouTube视频向公众展示。

项目成果

A neuromorphic SLAM architecture using gated-memristive synapses

• DOI: 10.1016/j.neucom.2019.09.098
• 发表时间: 2020-03
• 期刊: Neurocomputing
• 影响因子: 6
• 作者: Alexander Jones;A. Rush;Cory E. Merkel;Eric Herrmann;A. Jacob;Clare D. Thiem;R. Jha

Development of a Short-Term to Long-Term Supervised Spiking Neural Network Processor

短期到长期监督尖峰神经网络处理器的开发

• DOI: 10.1109/tvlsi.2020.3013810
• 发表时间: 2020
• 期刊: IEEE Transactions on Very Large Scale Integration (VLSI
• 作者: Bailey, Tony James;Ford, Andrew J.;Barve, Siddharth;Wells, Jacob;Jha, Rashmi
• 通讯作者: Jha, Rashmi