🕸️ Neural Networks

Welcome to the main page of my course, Neural Networks.

../../../_images/cover_image9.jpg

Believe it or not, it’s an AI era now!

Why neural networks?

Deep learning and neural networks, one of the hottest and most powerful subjects in machine learning/AI, has started revolutions in almost all fields, of course, including applied/computational mathematics.

Riding on this tide is never a bad choice. It sparks thoughts and guide the way for your research regardless of its topic. And, at the very least, having an item named “master neural networks theory and applications” listed on your CV makes it twinkling.

Goals

  • This course will first equip you with knowledge on fundamental theories, basic algorithms and some advanced algorithms of neural networks.

  • The point is not to know those introduced in this course; it enables you to understand any neural networks in the literature.

  • Not just theoretical studies, it also trains your hands-on skills to make sure that, during or after the course, you can always exam your ideas by coding them up.

Lecture contents

The following contents will be lectured.

  • Introduction to Artificial Neuron

  • Neural Network Architectures and Training Processes

  • (Multilayer) Perceptron (MLP) Network

  • Radial Basis Function (RBF) Networks

  • Recurrent Neural Networks (RNN)

  • Convolutional Neural Networks (CNN)

  • Physics-Informed Neural Networks (PINN)

More or less contents could be lectured depending on the lecture hours. Typically, it takes no less than 32 lecture hours, and every 16 lecture hours values 1 credit.

Hint

To follow this course, you should be aware of following aspects.

  • The course material will be fully in English. The lectures will be given in a bilingual (Mandarin/English) form.

  • The hands-on programming exercises and the final project will be done with PyTorch of Python. It is not a big deal if you do not have any Python programming experience yet. You can absolutely learn it during the course. It is very easy.

  • To make sure your gain is real, the course will be given always in the same standard no matter it is elective or compulsory. So make no wish that you can pass it effortlessly (though it is not difficult also).

How to pass?

Project

You will carry on a “big assignment”, namely, the project, and produce a report on it at the last stage of this course, individually or in a group (depends on the particular projects and number of students in this course). A project score of 0-100 will be graded based on your report.

Some project topics will be provided by the lecturer (they will be listed at The project pool of this page).

You are encouraged to set up your own project topic based on, for example, your research or any other topics that interest you.

Note: You can discuss with each other, search the internet, ask, for example, chatgpt, etc., but you cannot directly copy, or request chatgpt to generate contents for you.

score: 0-100

Oral exam

With a project report, an oral exam ends the course. Please do not worry, the oral exam is not a real “exam”. It is a short open discussion on your report to check whether it is really “your” report. If it indeed is, the oral exam will be enjoyable; you do not need to prepare for it and will get a score 1 no matter what score the report is.

But if you did not do anything and copied a report somewhere, you cannot hide the fact and will be “rewarded” with score 0 for the oral exam.

score: 0 or 1

The final score will be the multiplication of the project score and the oral exam score. And you need a final score of at least 60 to pass the course. That is saying your report needs to be graded \geq 60 and your oral exam needs to be scored 1.

The project pool

The projects shown selectable (✅) can be selected for the ongoing course.

Project

Selectable

Member(s)

Project one

=1

A really really long project title to test the layout of the table.

\leq 3

✨ Customize a project by yourselves

\leq 4

⛔ These projects are not available for the ongoing course.

✨ Bonus! You will be rewarded with extra scores.

The member(s), for example,

  • =1 means you can only conduct this project individually,

  • \leq 3 means you can do it individually or in a team of at most 3 people.

🛎️ Attention: A project in principle can be selected by multiple teams (or individuals). But if a project is too heavily selected, the teams (or individuals) that selected it late may be asked to re-select another project.

Contact lecturer


🚀 This course keeps evolving. The nearest edit was on Oct 19, 2024 at 17:33 (+0800).

↩️ Back to 📖 TEACHING.

🏠️ Back to Math is cheap.