MS004320-基于神经网络的微波电路设计（全英文）

研究生课程开设申请表

开课院（系、所）：亚星会员登录       

课程申请开设类型： 新开     重开□     更名□（请在□内打勾，下同）

一、课程介绍（含教学目标、教学要求等）（300字以内）

《基于神经网络的微波电路设计》将介绍用于高频高速微波电路计算机辅助设计的人工神经网络方法，包括无源和有源器件和电路的建模方法，以及它们在更高层次的有线和无线电子系统设计和优化中的应用，是信息类专业的一门重要的选修课。神经网络是受人脑从观察中学习和通过抽象概括的能力启发的信息处理系统，目前被应用于高频高速微波电路和系统的计算机辅助设计。本课程主要讲解神经网络结构、神经网络训练算法、无源和有源器件知识型神经网络模型、基于神经网络的微波电路快速仿真和优化设计等四部分内容，在培养学生独立思考、文献调研、分析问题、解决问题的能力等方面起到重要作用。同时，课程将实施系列改革，将思政教育元素和思政育人功能融入到课堂教学实践，培养学生的家国情怀、社会责任和工匠精神等。

二、教学大纲（含章节目录）：（可附页）

1.绪论和神经网络结构：（4学时）

了解神经网络模型的发展历史和现状，掌握神经网络结构的基础知识。

介绍计算机辅助设计技术的发展历史，着重介绍应用电磁仿真技术在微波设计优化过程中的局限性。通过一系列的案例介绍， 让学生了解到神经网络技术在加速微波器件仿真与优化设计中的作用。介绍几种典型神经网络的基本结构。

2.神经网络训练：（4学时）

网格型数据采样，自适应数据采样，自动模型结构生成，基于局部优化的神经网络训练算法，基于全局优化的神经网络训练算法。

3.神经网络在器件建模和电路仿真与优化中的应用：（4学时）

无源器件建模、有源非线性器件建模，基于神经网络的快速微波电路仿真与不确定性分析，神经网络辅助的优化设计，成品率驱动的优化设计。

4.知识型神经网络和其他先进网络结构与训练算法：（4学时）

几种典型的知识型神经网络结构，空间映射技术基本原理，深度神经网络结构及其训练算法，结合局部优化与全局优化的训练算法，基于并行计算的神经网络模型训练。

5.专题实验（6学时）

采用当下主流的编程平台（Matlab、C++或Python）进行编程实验，可由最多3人组成小组完成，一共3次，内容分别为：

1）MLP网络构建与模型训练实验

2）有源微波器件建模实验

3）无源微波器件建模实验

要求完成：

• 完整的编程代码

• 详细的编程实验报告word文档

6  研讨（10学时）

针对相关编程实验进行研讨：

1）MLP网络构建与模型训练实验（2学时）

2）有源微波器件建模实验（2学时）

3）无源微波器件建模实验（2学时）

针对当下的主流软件/算法进行研讨：

4）最新的电磁领域计算机辅助设计软件（2学时）

5）训练算法与优化技术的最新进展（2学时）

要求完成：

• 研讨ppt文稿

• 分组演讲并回答研讨提问

三、教学周历

注：1.以上一、二、三项内容将作为中文教学大纲，在研究生院中文网页上公布，四、五内容将保存在研究生院。2.开课学期为：春季、秋季或春秋季。3.授课语言为：汉语、英语或双语教学。4.适用学科范围为：公共，一级，二级，三级。5.实践环节为：实验、调研、研究报告等。6.教学方式为：讲课、讨论、实验等。7.学位课程考试必须是笔试。8.课件地址指在网络上已经有的课程课件地址。9.主讲教师简介主要为基本信息（出生年月、性别、学历学位、专业职称等）、研究方向、教学与科研成果，以100至500字为宜。

四、主讲教师简介：

张嘉男，教授、博导、东南大学青年首席教授，入选国家高层次青年人才、中央高校优秀青年团队、东南大学“紫金青年学者”。长期从事智能电磁计算、电磁超表面优化设计、量子计算在电磁问题中的应用等方向的研究。发表领域知名期刊/会议论文70余篇，主持国家自然科学基金优秀青年（海外）、江苏省自然科学基金青年基金、华为横向合作等项目，同时参与国家重点研发计划、JWKJW重点项目等国家级项目。

五、任课教师信息（包括主讲教师）：

Application Form For Opening Graduate Courses

School (Department/Institute)：

Course Type: New Open   Reopen □   Rename □（Please tick in □, the same below）

1. Course Introduction (including teaching goals and requirements) within 300 words:

This course will introduce artificial neural network methods for computer-aided design of high-frequency and high-speed microwave circuits, including modeling methods for passive and active devices and circuits, as well as their applications in higher-level wired and wireless electronic system design and optimization. It is an important elective course for information majors. Neural networks are information processing systems inspired by the human brain's ability to learn from observation and summarize through abstraction. They are currently applied in computer-aided design of high-frequency and high-speed microwave circuits and systems. This course mainly covers four parts: neural network structure, neural network training algorithms, knowledge-based neural network models for passive and active devices, and fast simulation and optimization design of microwave circuits based on neural networks. It plays an important role in cultivating students' abilities in independent thinking, literature research, problem analysis, and problem-solving. At the same time, the curriculum will implement a series of reforms, integrating elements of ideological and political education and the function of ideological and political education into classroom teaching practice, cultivating students' patriotism, social responsibility, and craftsmanship spirit.

2. Teaching Syllabus (including the content of chapters and sections. A sheet can be attached):

1. Introduction and Neural Network Structure: (4 class hours)

Understand the development history and current situation of neural network models, and master the basic knowledge of neural network structures.

Introduce the development history of computer-aided design technology, with a focus on the limitations of applying electromagnetic simulation technology in microwave design optimization. Through a series of case studies, students will understand the role of neural network technology in accelerating microwave device simulation and optimization design. Introduce the basic structures of several typical neural networks.

2. Neural network training: (4 class hours)

Grid based data sampling, adaptive data sampling, automatic model structure generation, neural network training algorithm based on local optimization, and neural network training algorithm based on global optimization.

3. Application of neural networks in device modeling, circuit simulation and optimization: (4 class hours)

Passive device modeling, active nonlinear device modeling, fast microwave circuit simulation and uncertainty analysis based on neural networks, optimization design assisted by neural networks, and yield driven optimization design.

4. Knowledge based neural networks and other advanced network structures and training algorithms: (4 class hours)

Several typical knowledge-based neural network structures, basic principles of spatial mapping technology, deep neural network structures and their training algorithms, combined with local and global optimization training algorithms, and neural network model training based on parallel computing.

5. Special Experiment (6 class hours)

Adopting current mainstream programming platforms (Matlab C++or Python programming experiments can be conducted in groups of up to 3 people, for a total of 3 times. The content includes:

1) MLP Network Construction and Model Training Experiment

2) Active microwave device modeling experiment

3) Passive microwave device modeling experiment

Required completion:

• Complete programming code

• Detailed programming experiment report Word document

6. Seminars (10 class hours)

Discussion on relevant programming experiments:

1) MLP Network Construction and Model Training Experiment (2 class hours)

2) Active Microwave Device Modeling Experiment (2 class hours)

3) Passive Microwave Device Modeling Experiment (2 class hours)

Conduct discussions on current mainstream software/algorithms:

4) The latest computer-aided design software in the electromagnetic field (2 class hours)

5) The latest progress in training algorithms and optimization techniques (2 class hours)

Required completion:

• Discuss the PowerPoint presentation

• Group presentations and answering discussion questions

3. Teaching Schedule:

 Note: 1.Above one, two, and three items are used as teaching Syllabus in Chinese and announced on the Chinese website of Graduate School. The four and five items are preserved in Graduate School.

 2. Course terms: Spring, Autumn , and Spring-Autumn term.   

 3. The teaching languages for courses: Chinese, English or Chinese-English.  

 4. Applicable range of discipline: public, first-class discipline, second-class discipline, and third-class discipline.  

 5. Practice includes: experiment, investigation, research report, etc.  

 6. Teaching methods: lecture, seminar, practice, etc.  

 7. Examination for degree courses must be in paper.  

 8. Teaching material websites are those which have already been announced.  

 9. Brief introduction of chief lecturer should include: personal information (date of birth, gender, degree achieved, professional title), research direction, teaching and research achievements. (within 100-500 words)  

4. Brief Introduction of Chief lecturer:

Jia Nan Zhang received the B.Eng. degree in Tianjin University, Tianjin, China, in 2013. He received the Ph.D. degree in the School of Microelectronics at Tianjin University, Tianjin, China, and the Department of Electronics at Carleton University, Ottawa, ON, Canada, in 2020.  

From 2020 to 2022, he was a Post-Doctoral Research Associate in the Department of Electronics at Carleton University, Ottawa, ON, Canada. In January 2022, he joined the State Key Laboratory of Millimeter Waves at Southeast University, Nanjing, China, where he is currently a Professor. He has authored/co-authored over 70 publications in prestigious microwave journals/conferences. His research interests include machine-learning approaches to microwave design, surrogate modeling and surrogate-assisted optimization, finite element analysis in EM, and quantum computing with applications to EM problems.

5. Lecturer Information (include chief lecturer)