Core Machine Learning Practices

This unit develops practical skills for preparing data, training and evaluating core ML models, and maintaining secure, reproducible workflows. Students will work with common data structures and notebooks, apply statistical and probabilistic reasoning to measure generalization, and practice algorithm implementation, explainability, and ethical safeguards in mid-scale settings.

Learning Objectives

• Analyze datasets and apply appropriate Python data structures (sets, lists, tuples, dictionaries) and notebook techniques (magic functions, indexing) to ingest, clean, and organize data for machine learning experiments
• Apply probability and statistical concepts to estimate out-of-sample error, design and execute correct cross-validation, and quantify bias–variance tradeoffs to detect and mitigate overfitting
• Demonstrate implementation, tuning, and comparison of core algorithms (greedy classification trees, distance-based clustering, ensembles, and basic neural networks), including optimization approaches for larger datasets and secure handling of secrets and sanitized inputs
• Evaluate model explainability, reproducibility, and ethical implications by interpreting algorithmic decisions, documenting notebook workflows, and proposing mitigations for AI risks and future-technology pitfalls