Which type of learning relies on historical data with labeled outcomes?

Supervised learning is a type of machine learning that relies on historical data that includes both input data and the corresponding correct output or labeled outcomes. In this approach, a model is trained on a dataset where each example has known labels, allowing it to learn the relationship between the inputs and the outputs. The goal is for the model to be able to make accurate predictions or classifications on new, unseen data based on the patterns it learned during training.

For instance, in a supervised learning task such as email classification, the model is trained on a dataset of emails that have been labeled as either "spam" or "not spam." By learning from these labeled examples, the model can then classify new emails into the appropriate category.

In contrast, other types of learning do not focus on labeled outcomes. Unsupervised learning analyzes data without labeled responses, seeking to identify patterns or groupings. Reinforcement learning involves agents learning through trial and error, receiving feedback from their actions in an environment rather than explicit labels for each input. Deep learning is a subset of machine learning that uses neural networks, which can be applied to both supervised and unsupervised learning, but its categorization doesn't specifically refer to the training on labeled outcomes.