midQ-Learning
What is reward shaping?
Updated May 17, 2026
Short answer
Reward shaping modifies rewards to speed up learning.
Deep explanation
It adds intermediate rewards to guide agent toward desired behavior without changing optimal policy.
Real-world example
Used in navigation tasks to encourage shorter paths.
Common mistakes
- Changing optimal policy unintentionally.
Follow-up questions
- What is potential-based shaping?
- Why is it needed?