Reward engineering — practical guidelines.
mediumAnswer
- (1) Start with sparse task reward (1 for success, 0 otherwise) — cleanest signal.
- (2) If too sparse, add potential-based shaping (safe).
- (3) Test agent for reward hacking: does it optimize what you meant?
- (4) Include safety / constraint penalties.
- (5) Normalize magnitudes across components.
- (6) Log all reward components separately for debugging.
- (7) Iterate.
- Interview red flag: agent 'learns' but does something weird → reward is wrong.
Check yourself — multiple choice
- Random
- Start sparse task reward → add potential-based shaping → test hacking + safety penalties + normalize magnitudes + log components; iterate — weird behavior = reward wrong
- Just guess
- Not real
Reward engineering: sparse first + safe shaping + hack test + normalize + log.
#reward-design#engineering
Practise Reinforcement Learning
214 interview questions in this topic.