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

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.

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.