Machine Learning

COMP 4211
The Hong Kong University of Science and Technology
Newt Nguyen

Course description: This course provides a comprehensive coverage of the machine learning field. There are 2 parts of this course: technical part and theoretical part. For the technical part, it will introduce necessary Python frameworks and OOP-related concepts for machine learning programming. For the theoretical part, it covers traditional machine learning and newly-emergent topics in large language models, computer vision, natural language processing, and some selected advanced topics.

Time and Location

Lectures: TBD

Some Reference books (will be updated as we go)

Participants should be comfortable with probability, linear algebra, and programming in Python. Helpful references:

Lecture Schedule

Schedules are tentative and subject to change. Slides will be posted here when available.

Technical Part 1 (programming & frameworks)

Python, OOP, and tooling for machine learning implementation.

SlidesDateTopicReadingsAssignments
COMP 2211 Python Basics
Extra NumPy
Week 1Matrix Programming Basics  
Project 1Week 1-2Project 1  
PandasWeek 2Pandas  
OOP for PythonWeek 2OOP for Python  
Scikit-Learn ProjectWeek 2
7th June 2026
Scikit-Learn Project  
K Nearest NeighborsWeek 2K Nearest Neighbors  
K Means ClusteringWeek 2K Means Clustering  
Project 2Week 2Project 2  
 Week 2Introduction to Keras  
 Week 3Introduction to PyTorch  
 Week 4Introduction to Hugging Face  

Theoretical Part 1 (ML concepts & applications)

Foundations, classical algorithms, and selected modern topics (LLM, CV, NLP).

SlidesDateTopicReadingsAssignments
  IntroductionCS229 Notes 
Linear Regression
Written Notes
Video Link
7th June 2026Linear Regression  
Logistic Regression Logistic Regression  
Feedforward Neural Networks Feedforward Neural Networks  
Deep Neural Networks Deep Neural Networks  
  Stochastic Gradient Descent  
  Automatic Differentiation  
  Convolutional Neural Networks  
  Recurrent Neural Networks  
  Autoencoders  
  Reinforcement Learning  

Theoretical Part 2 (LLM & GPU Performance)

SlidesDateTopicReadingsAssignments
  Language Models Architecture  
  LLM Pretraining  
  Nvidia GPU Performance  
  Nvidia Collective Communication Library  
  Data and Pipeline Parallel Training  
  Tensor Parallelism and ZeRO Optimizer  
  Mixture-of-Experts and Sequence Parallelism  
  Generative Inference Overview  

Theoretical Part 3 (Computer Vision)

SlidesDateTopicReadingsAssignments
Linear Classifier Image Classification with Linear Classifier  

Paper Discussion (will update as we go)

Paper LinkPaper NameTopic
Nielsen et al. (2026)Learning to Orchestrate Agents in Natural Language with the ConductorLLM Conductors
Wang et al. (2024)Mixture-of-Agents Enhance Large Language Model CapabilitiesLLM Conductors
Yue et al. (2025)MasRouter: Learning to Route LLM for Multi-Agent SystemLLM Conductors
Madaan et al. (2023)Self-Refine: Iterative Refinement with Self-FeedbackLLM Conductors
Du et al. (2023)Improving Factuality and Reasoning in Language Models through Multiagent DebateLLM Conductors
Zheng et al. (2025)Survey on LLM for Scientific DiscoveryLLM for Scientific Discovery
Jiang et al. (2025)AIDE: AI-Driven Exploration in the Space of CodeAI4AI
LLM Reasoning
Liu et al. (2025)ML-MasterAI4AI
LLM Reasoning
Zhu et al. (2026)ML-Master 2.0AI4AI
LLM Reasoning
Dosovitskiy et al. (2021)An Image is Worth 16x16 Words: Transformers for Image Recognition at ScaleComputer Vision
Li et al. (2025)WebSailorAI Research Agent