This course provides an introductory overview of modern techniques for learning from data. Students will learn the theories and applications of machine learning and deep learning. Topics include various neural network architectures – feedforward neural networks, convolutional neural networks, recurrent neural networks, transformers, large language models, graph neural networks, autoencoders, and deep generative models – along with optimization and regularization techniques for neural networks.
* This course introduces advances in deep learning, assuming that you have a basic understanding of machine learning and the required fundamental mathematics (Linear Algebra, Probability and Statistics). I will briefly review some prerequisite concepts, but not comprehensively.
** This course does not cover implementation using deep learning frameworks such as PyTorch or TensorFlow. But, you will need to use these frameworks using your own computational resources for your term project.
*** If you do not meet these prerequisites, I strongly recommend that you drop the course.
Class Time: Monday, 15:00~17:45
Location: 26513 (Engineering Building 2)
Language: Korean
Prof. Seokho Kang
Office: 27408B (Engineering Building 2)
E-mail: s.kang@skku.edu
Office Hours: by appointment
Ms. Jinju Park
Office: 27407 - Data Mining Lab. (Engineering Building 2)
E-mail: apfhsk777 at naver.com
Ian Goodfellow, Yoshua Bengio & Aaron Courville, Deep Learning, MIT Press, 2016. (https://www.deeplearningbook.org/ )
Aston Zhang, Zachary C. Lipton, Mu Li, Dive into Deep Learning, Cambridge University Press, 2023. (https://d2l.ai/)
Trevor Hastie, Robert Tibshirani & Jerome Friedman, The Elements of Statistical Learning: Data Mining, Inference, and Prediction (2nd ed.), Springer, 2009. (https://web.stanford.edu/~hastie/ElemStatLearn/)
PyTorch Tutorials (https://pytorch.org/tutorials)
Attendance (10%)
Presentation (10%)
Term Project (10%)
Mid-term Exam (30%)
Final Exam (40%)
Total (100% + a)
Syllabus [download]
Course Introduction [download]
1. Machine Learning Basics [download]
2. Deep Neural Networks [download]
3. Optimization [download]
4. Regularization [download]
5. Practical Methodology [download]
6. Convolutional Neural Networks [download]
7. Recurrent Neural Networks [download]
8. Transformers [download]
9. Large Language Models [download]
10. Graph Neural Networks [download]
11. Autoencoders [download]
12. Deep Generative Models [download]
13. Further Topics in Deep Learning [download]
All assignments should be submitted to icampus by midnight of the due date. Late submissions will NOT be accepted.
[A1] Self-Introduction
[A2] Term Project Proposal
[A3] Term Project Presentation
For the term project, students are required to participate in a Kaggle (https://kaggle.com/) or Dacon (https://dacon.io/) competition with a submission deadline no earlier than November 2025. The selected competition should be related to the concepts covered in the course. Students may work individually or in teams of up to three members. Evaluation will be based on the team’s public ranking on the competition leaderboard, as well as the overall completeness and quality of the final report. The report must clearly describe the problem being addressed, data preprocessing steps, model architecture, learning objectives, training strategies, and evaluation results. It should also include reflections on what worked well, challenges encountered, and suggestions for future improvements. Detailed deadlines for the proposal submission, presentation, and final report will be announced separately.
Students are invited to give a voluntary 10-minute presentation in class. The goal is to share recent advances in deep learning by presenting journal papers or conference proceedings (not necessarily their own), published within the last five years, that are related to topics covered in the course. The presentation topic should be relevant to the lecture topic of the same week. Students who wish to present must schedule it and obtain approval from the instructor at least two weeks in advance via email. Presentation materials must be submitted to the instructor via email at least one day before the presentation. Well-prepared and thoughtful presentations may be eligible for extra credit.
Students are responsible for maintaining high standards of academic integrity in all of their class activities. Cheating or plagiarism in any form will not be tolerated. Any violation of academic integrity is a serious offense and is therefore subject to an appropriate sanction or penalty.