What is one characteristic of unsupervised learning?

Unsupervised learning is characterized by its ability to identify hidden patterns or intrinsic structures in data without the need for labeled input. This approach allows algorithms to explore the data, find groupings or clusters, and discover relationships that may not have been evident or explicitly identified by a human. The essence of unsupervised learning lies in its exploratory nature, enabling it to derive insights from data sets where no predefined categories or labels exist.

The other options, while relevant to machine learning as a whole, do not accurately describe the nature of unsupervised learning. For instance, the idea that it requires labeled data for training directly contradicts the fundamental principle of unsupervised learning. Also, the focus on prediction and classification suggests a supervised approach, where models are trained on labeled data to predict outcomes, rather than uncovering patterns in unlabeled data. Lastly, stating that it is limited to supervised tasks completely misunderstands the distinction between supervised and unsupervised learning algorithms. Hence, the correct identification of finding hidden patterns without human guidance encapsulates the core attribute of unsupervised learning.