Course Introduction

Xijia Liu

Department of Statistics, Umeå University

About me

Introduction

  • Main topic: Non-linear Models
  • Artificial Neural Networks, Deep Learning basic.
  • Expected Learning outcomes: \[ \text{Understanding} > \text{Implementation} >0 \]
  • 3 credits \(\approx\) 4 weeks
  • 4 weeks \(+\) Pre knowledge from Part I \(=\) A challenge
  • So, my strategy (plan (design)) is…

Course Design

There are many ways to introduce and explain neural networks, most of which require learners to have a certain level of prior knowledge.

Course Design

However, how to quickly understand deep learning based on neural network models within a limited time, building on the first part of the course, is a challenge.

Course Design

Here, I will use the most cost-effective approach to help you deeply understand neural network models. No free lunches. We first need to take some time to build a bridge.

Course Design

To build this bridge, we will introduce the concept of feature extraction in the first lecture and present the most fundamental data-driven solution, PCA.

Course Design

Then, in the second lecture, we will revisit PCA from a machine learning perspective using a practical example.

Course Design

I believe that once you fully understand the content of the first two lectures, you will smoothly grasp ANN models and the core concepts of deep learning.

Teaching Methods

  • Online course \(\to\) Flipped classroom method
  • Read and study learning materials
  • Two Zoom meetings. In each meeting,
    1. Discussion and Q/A \(\approx 1\) hour.
    2. Lab session \(\approx 2\) hours.
  • Communication: Canvas messages

Learning Materials

  • Textbook: Introduction to statistical learning with application in R
  • My lecture notes and materials on Yggdrasil

Yggdrasil is the “world tree” in Norse mythology. My Yggdrasil is about data science.

Exam (Mandatory)

  • Project study
  • Maximum two students
  • Suggested problems or topics you are interested in
  • Oral exam:
    1. Individual meeting on Zoom
    2. Discuss your project study
    3. Other course related questions
  • Timetable: check it on Canvas