Foundations of Deep Learning

This unit introduces core deep learning and machine learning concepts, essential mathematics, and practical tools (Python and common frameworks) for building and training simple neural networks. Students will explore basic architectures (perceptron, feedforward, CNN, RNN), data and task types, and the real-world considerations involved in deploying and managing deep learning projects to prepare for more advanced units.

Learning Objectives

• Analyze foundational terminology and distinctions among machine learning, deep learning, and common model families (perceptron, feedforward, CNN, RNN, generative/adversarial) and how they map to problem types
• Apply Python and a deep learning framework to implement and train simple models (perceptron and small neural networks) on example datasets such as character recognition and image classification
• Demonstrate the core mathematical operations used in neural networks by combining variables, computing forward passes and weight updates, and explaining gradients for simple examples
• Evaluate dataset characteristics and classification/regression tasks to select appropriate model architectures and justify considerations for real-world use, team roles, and ethical/deployment implications