Model Training & Adaptation
Policy (reinforcement learning)
The rule or learned approach an AI uses to choose its next action.
Example
An agent selects a move based on the current game state.
Why people use it
It describes the decision rule an agent follows when choosing what to do.
What you'll hear
“What guides its next move?”
What this means for you
Identify which meaning of policy is intended in a technical discussion.
Can you control it?
Developer-only
The people building or running the AI choose this setup. An everyday user generally needs their help to change how this part works.
Common questions
- Is a reinforcement-learning policy the same as an organizational AI policy?
- No. One chooses actions; the other states rules for people and systems.
- Can a policy include randomness?
- Yes. It may choose among several actions with different chances rather than always choosing one.
- Can the policy change during learning?
- Yes. Feedback can change which actions the agent tends to choose.