Machine Learning (CECS 456)

Course Description

This course introduces the foundations of machine learning through the lens of predictive modeling and function approximation. Students will learn how to build models for regression and classification using both discriminative and generative approaches, and will explore a range of modeling techniques including neural networks, training dynamics, regularization, and performance evaluation.

The course also introduces advanced topics such as convolutional architectures and transformers (if time permits). Emphasis is placed on understanding model assumptions, deriving algorithms, and analyzing the mathematical and statistical properties of learning systems. Students will implement learning algorithms and conduct experiments to evaluate models in practical settings.

Learning Outcomes

By the end of this course, students will be able to:

• Formulate regression and classification tasks using probabilistic modeling.
• Implement and compare linear, non-linear, and kernel-based models.
• Analyze model behavior and generalization using training curves and regularization.
• Derive and compute gradients for model training.
• (Optionally) Describe and apply deep learning architectures such as CNNs and transformers.

Modules

1. Foundations of Learning & Inference

• Learning vs. inference
• Discriminative vs. generative models
• Regression and classification as canonical problems
• Supervised learning overview and evaluation

2. Regression Models

• Linear and non-linear regression models
• Sparse regression and relevance vector regression
• Gaussian process regression
• The kernel trick

3. Classification Models

• Logistic regression and multi-class extensions
• Non-linear and relevance vector classification
• Gaussian process classification
• Incremental fitting: boosting and trees

4. Neural Networks & Training Dynamics

• Shallow and deep neural networks
• Loss functions, initialization, and fitting models
• Computing gradients and backpropagation
• Regularization and performance evaluation

5. Convolutional Networks & Vision

• Convolutional architectures
• Image generation models and applications

6. Transformers and LLMs (Optional)

• Attention mechanisms and transformer design
• Scaling and pretraining in large language models

Evaluation Components

Extra credit opportunities may be offered for solving advanced problems or conducting research-based explorations.

Grading Policy

Learning Resources

There is no required textbook for this course; readings will be assigned from open educational resources. However, the following books are highly recommended for students who wish to build a deeper understanding:

• Understanding Deep Learning by Simon J.D. Prince – A modern and accessible introduction to deep learning, covering both core concepts and recent advances.
• Probabilistic Machine Learning: An Introduction by Kevin P. Murphy – A comprehensive treatment of machine learning with a focus on probabilistic modeling.
• Deep Learning: Foundations and Concepts by Christopher M. Bishop and Hugh Bishop – A foundational text introducing the key ideas and methods in deep learning.

Special Accommodations

If you require accommodations due to a documented disability or other conditions, please notify the instructor early and contact the Bob Murphy Access Center (BMAC).

Students facing food or housing insecurity should connect with the Student Emergency Intervention & Wellness Program or Basic Needs Program. For mental health support, visit Counseling and Psychological Services (CAPS).

Academic Integrity

All students must adhere to the CSULB Academic Integrity Policy. Plagiarism, cheating, or misuse of generative AI tools are strictly prohibited. Work submitted must be original and appropriately cited. Violations will result in penalties including zero credit and potential grade reductions.

Disclaimer

This syllabus is subject to change. Students are expected to check Canvas regularly for announcements, updates to the schedule, and policy adjustments.