Unit 3: AI

This unit, focused on artificial intelligence, explores some of the ideas in New York's Computer Science and Digital Fluency Computational Thinking and Impacts of Computing standards.

Students build machine learning systems from the ground up. In Estimation, they fit a line to data by hand before training a real regression model, meeting the vocabulary—model, parameters, loss, training—that every later lab builds on. In Classification: Features, they build a spam detector, progressing from hand-written rules to hand-designed features to a bag-of-words model, and learn to evaluate a classifier honestly. In Classification: Neural, they take on handwritten digit recognition, discovering why hand-designed features fall short and building multi-layer perceptrons and convolutional neural networks from scratch. In Reinforcement Learning, they train an agent to play games without being told the rules, implementing Q-learning by hand before running their own training experiments. Throughout, technical work is paired with questions about bias, causation, and the ethics of building and deploying predictive systems.