Which statement correctly describes learning paradigms: supervised, unsupervised, and reinforcement learning?

• A. Unsupervised learns via rewards.
• B. Supervised uses unlabeled data.
• C. Reinforcement learning requires labeled data.
• D. Supervised uses labeled data; unsupervised finds structure; reinforcement learns via actions and rewards.

Answer: (D)

Three learning paradigms differ in the data and feedback they rely on. In supervised learning, you train on input-output pairs where the true labels guide learning through a loss that compares predictions to those labels. In unsupervised learning, there are no labels; the goal is to discover structure in the data, such as clusters or compact representations. In reinforcement learning, an agent interacts with an environment, taking actions, receiving rewards, and learning a policy that aims to maximize cumulative reward over time. The statement that supervised uses labeled data; unsupervised finds structure; reinforcement learns via actions and rewards correctly captures these distinctions. The other options mix up where labels and rewards come into play—rewards drive reinforcement, not unsupervised; labeled data is not a requirement for unsupervised; and reinforcement does not rely on labeled data in the same way as supervised learning.