1
课程详述
COURSE SPECIFICATION
以下课程信息可能根据实际授课需要或在课程检讨之后产生变动。如对课程有任何疑问,请联
系授课教师。
The course information as follows may be subject to change, either during the session because of unforeseen
circumstances, or following review of the course at the end of the session. Queries about the course should be
directed to the course instructor.
1.
课程名称 Course Title
An Introduction of Machine Learning and EDA
机器学习及电子设计自动化概论
2.
授课院系
Originating Department
深港微电子学院 School of Microelectronics
3.
课程编号
Course Code
SMES201
4.
课程学分 Credit Value
1
5.
课程类别
Course Type
专业选修课 Major Elective Courses
6.
授课学期
Semester
夏季 Summer
7.
授课语言
Teaching Language
英文 English
8.
他授课教师)
Instructor(s), Affiliation&
Contact
For team teaching, please list
all instructors
陈全(助理教授)
深港微电子学院
邮箱:chenq3@sustech.edu.cn
办公室:第一教学楼 101
Quan CHEN(Assistant Professor)
School of Microeletronics
Emailchenq3@sustech.edu.cn
OfficeRoom101,Teaching Building 1
9.
/
方式
Tutor/TA(s), Contact
NA
10.
选课人数限额(不填)
Maximum Enrolment
Optional
2
授课方式
Delivery Method
习题/辅导/讨论
Tutorials
实验/实习
Lab/Practical
其它(请具体注明)
OtherPlease specify
总学时
Total
11.
学时数
Credit Hours
12.
先修课程、其它学习要求
Pre-requisites or Other
Academic Requirements
13.
后续课程、其它学习规划
Courses for which this course
is a pre-requisite
14.
其它要求修读本课程的学系
Cross-listing Dept.
教学大纲及教学日历 SYLLABUS
15.
教学目标 Course Objectives
本课程将介绍以下知识点:机器学习的基本概念和基础知识,一种流行的深度学习框架和一些可以通过深度学习解决的简
单任务,机器学习的可能研究方向, AI EDA 硬件实现的基础知识,教授电路仿真的基本模型和算法,AI 硬件和 EDA
未来主题。
Introduce the basic concepts and fundamentals of machine learning.
Introduce a popular deep learning framework and some simple tasks that can be solved with deep learning.
Introduce possible research directions of machine learning.
Introduce fundamental knowledge of hardware realization of AI & EDA.
Teach basic models and algorithms for circuit simulation.
Introduce future topics in AI hardware & EDA.
16.
预达学习成果 Learning Outcomes
通过本课程的学习,学生可以了解机器学习和深度学习的基本概念和机制、能够使用机器学习框架实现神经网络、了解深
度学习的可能研究方向、获取 AI EDA 设计流程的硬件基础知识、了解半导体器件和电路的基本模型、能够理解和实现
电路仿真的基本算法。
Understand the basic concepts and mechanisms of machine learning and deep learning.
Be able to implement neural networks with a machine learning framework.
Get to know possible research directions of deep learning.
Acquire basic knowledge of hardware for AI & EDA design flow
Understand basic models of semiconductor devices and circuits
Be able to understand and implement basic algorithms of circuit simulation
3
17.
课程内容及教学日历 (如授课语言以英文为主,则课程内容介绍可以用英文;如团队教学或模块教学,教学日历须注明
主讲人)
Course Contents (in Parts/Chapters/Sections/Weeks. Please notify name of instructor for course section(s), if
this is a team teaching or module course.)
Credit Hour 1-4: Fundamentals of Machine Learning 学时 1-4: 机器学习的基础知识
Day 1: Machine Learning basics: tasks (classification, regression), measurement (precision, recall, F-score), supervised
learning and example (linear regression), No Free Lunch Theorem
1 天:机器学习基础知识: 任务 (分类、回归)、测 (度、召回、f-)、监督学习和示例 (线性回)、无免费午餐
Day 2: Deep Neural Network (MLP), Training a DNN (gradient back-propagation)
2 : 深层神经网络 (MLP), 训练 DNN (梯度反向传播)
Day 3: Activation Function (sigmoid, ReLU->Leaky ReLU->PReLU), Loss function (MSE, Cross Entropy), Regularizer
(L1, L2), Optimizer (SGD)
3 : 激活函數 (sigmoid, relu > 泄漏的 ReLU -> PReLU), 损耗函数 (MSE, 交叉熵), 调节器 (L1, L2), 优化器 (SGD)
Day 4: Convolutional Neural Network (convolution, pooling, softmax), Recurrent Neural Network
4 : 卷积神经网络 (卷积、池化、softmax)、递归神经网络
Credit Hour 5-8: Deep Learning Practice 学时 5-8: 深度学习练习
Day 1: Python basics, Pytorch basics (tensors, autograd, nn module)
1 : Python 基础知识, Pytortorch 基础知识 (张量、自动研究生、nn 模块)
Day 2: Handwritten digits classification with CNN
2 : 基於 CNN 的手写数字分类
Day 3: Name Generation with RNN
3 : 基於 RNN 的名称生成
Day 4: Research Directions & showcase
第四天: 研究方向和應用展示
Credit Hour 9-12: Basic EDA Flow for VLSI 学时 9-12: VLSI 的基本 EDA 流程
Day 1: Introduction of VLSI (with emphasis on AI hardware)
1 : 超大规模集成电路介绍 (侧重人工智能硬件)
Day 2: Overview of EDA Flow
2 : EDA 流程概述
Day 3: Circuit Simulation Basics (SPICE, netlist, MNA, software)
4
3 : 电路仿真基础知识 (SPICE、网络列表、MNA、软件)
Day 4: Basic Semiconductor Devices Models (Drift-Diffusion model, compact models)
4 : 基本半导体器件模型 (漂移扩散模型, 集约模型)
Credit Hour 13-16: SPICE Circuit Simulation Methods & Algorithms 学时 13-16: SPICE 电路仿真方法和算法
Day 1: Linear Circuit Simulation (linear systems, Gaussian elimination, Iterative solvers)
1 : 线性电路仿真 (线性系统, 高斯消除, 迭代求解)
Day 2: Nonlinear Circuit Simulation (Newton’s iteration, time integration algorithms)
2 : 非线性电路仿真 (牛顿迭代, 时间积分算法)
Day 3: Physical Design I
3 : 物理设计 I
Day 4: Physical Design II
4 : 物理设计 II
18.
教材及其它参考资料 Textbook and Supplementary Readings
(教材)Deep Learning Book: http://www.deeplearningbook.org/
Neural Networks and Deep Learning: http://neuralnetworksanddeeplearning.com/
Pytorch Tutorials: https://pytorch.org/tutorials/
(教材)CMOS VLSI DESIGN: A Circuits and Systems Perspective”: http://pages.hmc.edu/harris/cmosvlsi/4e/index.html
SPICE Simulation Fundamentals: http://www.ni.com/en-us/innovations/white-papers/06/spice-simulation-
fundamentals.html
Electronic design automation (Wiki): https://en.wikipedia.org/wiki/Electronic_design_automation
课程评估 ASSESSMENT
19.
评估形式
Type of
Assessment
评估时间
Time
占考试总成绩百分比
% of final
score
违纪处罚
Penalty
备注
Notes
出勤 Attendance
30
课堂表现
Class
Performance
20
小测验
Quiz
课程项目 Projects
平时作业(两次)
Assignments
50
期中考试
Mid-Term Test
期末考试
Final Exam
期末报告
Final
5
Presentation
其它(可根据需
改写以上评估方
式)
Others (The
above may be
modified as
necessary)
20.
记分方式 GRADING SYSTEM
A. 十三级等级制 Letter Grading
B. 二级记分制(通/不通过) Pass/Fail Grading
课程审批 REVIEW AND APPROVAL
21.
本课程设置已经过以下责任人/员会审议通过
This Course has been approved by the following person or committee of authority