Preview
IB DP Computer Science — A4 Machine learning — Machine learning approaches | HL | A4.3.6
📋 What You Will Learn
- Describe the agent–environment interaction loop: states, actions, rewards and policies
- Explain the principle of cumulative reward and why an agent maximises it rather than the immediate reward
- Describe the exploration versus exploitation trade-off and the ε-greedy strategy that balances it
- Trace how an agent's value estimates update as it interacts with its environment
- Distinguish reinforcement learning from supervised and unsupervised learning
🐶 Why This Matters You do not train a puppy by handing it a labelled dataset of "correct sits". It tries actions, gets a treat when a good one happens, and gradually repeats what earned treats. Reinforcement learning trains software the same way — by trial, error and reward — and it is how systems learn to play games, control robots and manage traffic lights.
Where RL Sits Among the ML Approaches
- Supervised learning learns from labelled examples — the "right answer" is given for every training input.
- Unsupervised learning finds structure in