Which paradigm learns by trial and error through reward signals to improve its policy?

• A. Reinforcement Learning
• B. Supervised Learning
• C. Symbolic AI
• D. Expert Systems

Answer: (A)

Learning by trial and error with reward signals to improve behavior is reinforcement learning. In this approach, an agent interacts with an environment, chooses actions, receives feedback in the form of rewards, and updates its policy to maximize cumulative reward over time. The focus is on learning from feedback rather than from pre-labeled examples.

This differs from supervised learning, which relies on labeled input-output pairs to learn a direct mapping; and from symbolic AI and expert systems, which depend on explicitly encoded rules and reasoning over a knowledge base.

For instance, a robot learning to navigate a maze improves its policy by trying paths, receiving positive rewards for reaching the goal and negative signals for dead ends, and gradually choosing better routes.