What distinguishes supervised from unsupervised learning in NLP?

Supervised learning is characterized by its reliance on labeled data, where each training example is paired with an output label. This labeling process allows the model to learn to map input data to the correct output during training. In the context of natural language processing (NLP), supervised learning applies to tasks such as sentiment analysis or named entity recognition, where specific outcomes are predetermined and help guide the learning process.

On the other hand, unsupervised learning operates without labeled responses, allowing the algorithm to explore the structure of the data itself and identify patterns or groupings. This method is often utilized for tasks like topic modeling or clustering, where the aim is to uncover inherent structures within the data without predefined categories.

Thus, the distinction hinges fundamentally on the presence or absence of labeled data in training sets, which drives the learning process in supervised learning while leaving unsupervised learning to discover hidden structures independently.