| 课程号 |
00333754 |
学分 |
3 |
| 英文名称 |
Scientific Machine Learning: Blending Science with Data |
| 先修课程 |
无 |
| 中文简介 |
数据驱动技术可用于基于给定数据构建工程问题模型。然而，通常情况下，当模型应用于训练数据参数范围之外时，其性能往往不佳。为了获得更准确的预测结果，需要将问题的科学知识引入数学模型中。本课程将教授学生如何高效地运用现代人工智能和机器学习工具解决经典工程问题。课程重点在于将科学的工程领域知识与现代人工智能工具相结合，从而找到工程问题的最优解。此外，我们还将向学生介绍不确定性量化方法，以及如何利用这些技术来理解解对输入参数不确定性的敏感性。本课程的主要目标是使学生掌握数据驱动工具，并能够为简化的工程应用创建数学模型。课程将着重于将科学知识与数据驱动技术相结合，以确保严谨的科学原理融入模型之中。 |
| 英文简介 |
Data driven techniques can be used to find models for engineering problems based on given data. However, it is usually common that the derived models do not perform well when used outside the parameter range used in the training data. In order to obtain better predictions, scientific knowledge of the problem needs to be introduced into the mathematical model. In this course students will learn to efficiently apply modern day artificial intelligence and machine learning tools to classical problems in engineering. There will be emphasis on blending scientific engineering domain knowledge with modern AI tools to arrive at an optimal solutions to engineering problems. In addition, we will also introduce students to methodologies for uncertainty quantification and how these techniques can be used to understand the sensitivities of the solutions to uncertainties in input parameters. The main goal for this course is to arm students with data-driven tools that can be used to create mathematical models for simplified engineering applications. There will be an emphasis to blend scientific knowledge with data-driven techniques to ensure rigorous scientific principles is embedded into the model. |
| 开课院系 |
工学院 |
| 成绩记载方式 |
百分制 |
| 通识课所属系列 |
|
| 授课语言 |
英文 |
| 教材 |
|
| 参考书 |
|
| 教学大纲 |
The main goal for this course is to arm students with data-driven tools that can be used to create mathematical models for simplified engineering applications. There will be an emphasis to blend scientific knowledge with data-driven techniques to ensure rigorous scientific principles is embedded into the model.
Topics
1. Programming language is Julia (basic programs will be provided). Other programming languages, such as Matlab and Python, are also fine but are not supported
2. Basics of Uncertainty Quantification
3. Automatic Differentiation
4. Optimization
5. Introduction to Neural Network and regression.
6. Physics Informed Neural Network
7. Neural Ordinary Differential Equations
课堂授课
Assignment 1 20%
Assignment 2 30%
Final exam 40%
Attendance and discussion 10%
Total 100%
|
| 教学评估 |
|