Installs as a private draft. Your edits and self-improvement runs do not change the published bundle.
Add intermediate reward signals that guide learning without changing the optimal policy
Add intermediate reward signals that guide learning without changing the optimal policy
Sparse rewards make learning slow; shaping accelerates it
Reward Engineering > Reward Function Design
--- name: "Reward Shaping" version: "1.0.0" description: "Add intermediate reward signals that guide learning without changing the optimal policy" tags: - rl-skills-universe - "reward-engineering" - "reward-function-design" trigger: "reward-shaping" --- # Reward Shaping Add intermediate reward signals that guide learning without changing the optimal policy ## Why It Matters Sparse rewards make learning slow; shaping accelerates it ## Domain **Reward Engineering** > Reward Function Design