Which statement best describes unsupervised learning?

• A. Algorithm learns to maximise reward
• B. Algorithm uses input data without output data
• C. Algorithm relies on labeled data for supervision
• D. Algorithm uses feedback signals to adjust policy

Answer: (B)

Unsupervised learning is about discovering structure in data when there are no labeled outputs to guide the learning. The algorithm is given input features but no corresponding target values, so it tries to find patterns, groups, or compact representations on its own. This is why learning from input data without output data is the best description.

If labels were provided, it would be supervised learning, where the model learns to map inputs to known outputs. If learning hinged on interactions with an environment and adjusting decisions based on rewards, that would be reinforcement learning with feedback signals guiding a policy.