Introduction to Machine Learning Model Classes

Machine learning models are commonly divided into two classes: supervised and unsupervised. The key difference lies in the use of labels—supervised models train on data that includes explicit tags, while unsupervised models operate on data without any tags.

Supervised Learning

Supervised learning relies on labeled data, meaning each example comes with a tag such as a name, type, or number. Consider a restaurant owner who has historical records of bill amounts, tip amounts, and the order type (pickup or delivery). A supervised model ingests the bill amount as input and learns from past examples to predict the tip amount for future orders.

The learning process follows a classic optimization pattern: the model receives test data, generates a prediction, and then compares that prediction to the known training values. If the predicted value deviates significantly from the actual value, the difference is called “error.” The model iteratively adjusts its parameters to reduce this error until the predictions align closely with the real outcomes. As one concise statement puts it, “the key difference between the two is that with supervised models we have labels.”

Unsupervised Learning

Unsupervised learning works with unlabeled data, where no explicit tags are provided. Its goal is discovery—identifying natural groupings or patterns that exist within raw data. For example, a company might analyze employee tenure and income to see whether certain individuals fall into a “fast‑track” group. The algorithm examines the data and clusters employees based on similarity, revealing structures that were not previously labeled. This approach is summarized by the observation that “unsupervised problems are all about discovery about looking at the raw data and seeing if it naturally falls into groups.”

Relationship to Generative AI

Both supervised and unsupervised learning serve as foundational concepts for generative AI. Understanding how models learn from labeled examples or uncover hidden structures equips learners with the necessary background to grasp how generative systems create new content. As emphasized in the brief, “Understanding these Concepts is the foundation for your understanding of generative AI.”

Next Steps: Deep Learning

Having distinguished supervised from unsupervised learning, the logical progression is to explore deep learning, a specialized subset of machine learning that builds on these core ideas with layered neural networks. This deeper dive will further illuminate the mechanisms behind advanced AI applications.