EN813 706 Artificial Neural Networks

Semester 2/2567 (Nov 2024 - Mar 2025)

Course description

The class discusses development, motivation, background, key concepts, state-of-the-arts of artificial neural networks and related issues, e.g., basic machine learning, brain and biological neural networks, as time allows. Students are expected to actively participate in class and assignments.

---

Instructor

• Tatpong Katanyukul

Textbooks

• ธัชพงศ์ กตัญญูกุล การผจญภัยของการเรียนรู้ของเครื่องในโลกการรู้จำรูปแบบ 2023
• C. M. Bishop, Pattern Recognition and Machine Learning. Springer 2007
• Ian Goodfellow et al., Deep Learning. MIT Press 2016

---

Assessments (tentative)

• Participation: 10%
• Exercises/Homeworks/Assignments/Project: 40%
    • autolab.en.kku.ac.th
• Exams: 50%

---

Attendance policy

Students must attend no less than 80% of our classes. One attending the class less than 80% will be disqualified.

---

New tools or toys

  • CoE Jupyter Hub @ https://mozart.en.kku.ac.th:8443
  • CoE LLM @ https://bazille.en.kku.ac.th:8080

---

Main materials

• Orientation, introduction, and overview
    • Hands-on: prediction (Colab; Mozart)
    • Data
  
• Computation and numerical programming
    • Hands-on: vectorization, numpy, and matplotlib (Colab; Mozart)
    • Hands-on: numerical programming (Colab; Mozart)
• Optimization
    • Hands-on: optimization (Colab)
    • Hands-on: gradient descend algorithm (Colab)
  
• Introduction to machine learning
    • Hands-on: curve fitting (Colab)
  
• Generalization
    • Hands-on: overfitting and aux (Colab)
  
• Motivations/perceptrons
• Multi-layer perceptrons
    • Hands-on: mlp (from scratch)
    • Backpropagation (Colab)
    • Hands-on: sklearn (Colab)
  
• Practical tricks/applications
    • Input normalization; aux file; (Colab)
    • Classification (Colab)
    • Multiclass classification (Colab)
    • Application examples; aux file; (Colab)
      • Flood prediction (paper, org code, data prep:colab, random forest:colab)
  
• Deep learning
    • pytorch (1) (Mozart)
    • pytorch (2) (Mozart)
  
• Miscellaneous
    • PR Curve common mistake (Mozart)
  
• Selected topics
    • CNN motivation with keras (Colab)
    • CNN with keras (Colab)
    • Object detection (Colab)
    • Object detection: interactive version (Colab)
  
• Paper reading and discussion
    • AY2567/2 Vaswani et al., Attention is all you need (2017) ; Dig deeper: Hands-on , data and aux (Colab)
  
• Class project (tentative)

---

Additional materials

• Department compute server
• Introduction to Machine Learning.
•   • Hands-on: Sklearn with data and aux (colab)
•   • Hands-on: Regression (colab)
•   • Hands-on: Classification (colab)
•   • Clustering with K-Means
•     • Hands-on: K-Means, with example data Peruvian Girl Image (colab)
•   • Dimension reduction with PCA
•     • PCA on Eigenfaces
•     • Hands-on: PCA: Eigenfaces and Face Dataset (colab)
•     • PCA on Metabolomics
•     • Hands-on: PCA: Metabolomics and Liver Data (colab)
•   • Reinforcement Learning
•     • Hands-on: Reinforcement Learning (R Project)
• Overview to AI
•   • Agent
•   • Logic
•     • Hands-on: logic
•   • Handling uncertainty
•     • Hands-on: handling uncertainty
•   • Learning

---

Learn more

If you are interested in the development and state-of-the-art in the field, here's one good place to start with:

• Papers with Code
• Kaggle
• Latest development
  ** Conference on Neural Information Processing Systems (NeurIPs)
  ** International Conference on Learning Representations (ICLR)

---

Old class materials

• ANN 2023
• ANN 2022

---

Miscellaneous

COE Wiki

---

Academic Honesty

You are expected to do your own work to show understanding, skills, and what you have learned. All submitted works (including HOMEWORKS!) should be your own and ACADEMIC DISHONESTY IS NOT ALLOWED.
Academic dishonesty includes:

• Copying answers or codes;
• Copying words, ideas, codes, or other materials from another source without giving credit to the original author;
• Copying from your peers or seniors;
• Employing or letting another person to alter, revise, or edit your work, and then submitting the work as your own;
• Intentionally letting any of your peers to copy your work and submit as one's own;
• Submitting work automatically produced by an emerging tool (e.g., AI) as your work.

---

Last updated 2025 Mar 2nd.