ADSC 4720 - Data Mining in Applied Data Science

This course introduces students to the theory and practice of deep learning. It covers the architecture and operation of artificial neural networks, including training using gradient descent and backpropagation, and explores key challenges such as overfitting, vanishing gradients, and optimization. Students gain hands-on experience building and training models using PyTorch, and learn to apply advanced architectures such as Convolutional Neural Networks (CNNs), Recurrent Neural Networks (RNNs), and Transformers to real-world problems in vision and sequence modeling.

Learning Objectives¶

After completing this course, you will be able to:

• Describe the architecture and key components of artificial neural networks, including neurons, layers, activation functions, and loss functions.

• Explain how neural networks learn using gradient descent and backpropagation and identify challenges such as vanishing gradients and overfitting.

• Implement and train fully connected neural networks using PyTorch, including custom training loops, optimizers, and regularization techniques.

• Apply CNNs, RNNs, and Transformers models to image and sequence-based data problems.

• Evaluate deep learning models using performance metrics, validation techniques, and learning curves.

• Improve generalization using dropout, batch normalization, data augmentation, and learning rate scheduling.

• Interpret and explain model behavior using modern interpretability tools such as SHAP and feature visualization.

Schedule & Office Hours¶

• Tuesdays 11:30 - 12:15, OM 2652

• Wednesdays 1:30 PM - 2:45 PM, OM 1241

Office Hours: Wed 12:30 PM -13:30 PM in Clock Tower CT 409.

Communication¶

• Course-related questions: Please use the discussion forum on Moodle. This allows other students to benefit from the discussion.

• Individual-related questions: For academic concessions, extensions, or personal matters, please email me at lnguyen[at]tru[dot]ca.

• Response time: I aim to reply within 24 hours during normal working hours (9 AM – 5 PM, Mon–Fri).

Open sourced materials¶

• B1: Chollet, F., & Watson, M. (2025). Deep learning with Python, third edition. https://deeplearningwithpython.io

• B2: Zhang, A., Lipton, Z. C., Li, M., & Smola, A. J. (2023). Dive into deep learning. Cambridge University Press. https://d2l.ai/

Course Schedule¶

Assessment Overview¶

There will be two midterms and one final exam in this course.

The midterms are closed-book format, and they will take place on Moodle for the duration of 45 minutes. It will consist of a mix of multiple choice, fill in the blank, and open-ended questions.

The final exam is closed-book format, taking place on Moodle for the duration of 60 minutes. It will consist of a mix of multiple choice, fill in the blank, and open-ended questions.

The final exam will cover all the materials in the course.

Course policy¶

Attendance¶

A registered student who does not attend the first two events (e.g., lectures/labs/ etc.) of their course(s) and who has not made prior arrangements acceptable to the instructor(s) may, at the discretion of the instructor(s), be considered to have withdrawn from the course(s) and have their course registration(s) deleted.

• Arriving more than 5 minutes late will be recorded as absent.

• Missing more than 30% of class sessions will result in automatic failure of the course.

• Attendance accounts for 10% of your final grade.

• Each student gets three “free passes” for any reasons (e.g., illness, family matters, commuting issues) without penalty.

Please refer to TRU’s attendance policy. In addition, we will take attendance during class via Moodle’s QR code.

Academic Concessions¶

If circumstances (e.g., illness, family emergency, significant life event) may prevent you from meeting course requirements:

• Notify the instructor at least 24 hours before the deadline.

• Requests are considered case-by-case; you may be asked for documentation.

• Possible accommodations:

  • Deadline extensions

  • Alternative assessments

• Requests made after the deadline are usually refused.

Late Assignments¶

• Penalty: −25% per day, up to −75% total.

• After 3 days late, work is not accepted (grade = 0).

• Extensions are granted only for exceptional cases if requested before the deadline.

Accessbility¶

Students registered with the Accessibility Services who require accommodations must provide their Letter of Accommodation to the instructor as soon as possible. This letter will outline the necessary accommodations to ensure an equitable learning environment. Please ensure that this is done early in the term to facilitate timely arrangements.

Policy on the use of generative AI¶

I am a proponent of the responsible and ethical use of AI in education. You are welcome and encouraged to use any AI tools to support your learning process as you see fit. However, you are responsible for critically evaluating and verifying any AI-generated content you use. I also encourage you to explicitly acknowledge the use of AI in your work.

Suggested format template: AI Acknowledgement: This assignment was completed with assistance from [AI tool name, version, and provider]. The AI was used for [specific purpose, e.g., generating code snippets, summarizing readings, checking grammar]. All AI-generated content was reviewed, verified, and edited by me.

Please refer to TRU’s guideline on the use of generative AI for more information.