reward shaping
Reward shaping is a technique in reinforcement learning that modifies the reward signal received by an agent to make learning more efficient, often by providing additional intermediate rewards that guide the agent towards desired behaviors.
- Centralized Reward Agent for Knowledge Sharing and Transfer in Multi-Task Reinforcement Learning
- Fine-grained List-wise Alignment for Generative Medication Recommendation
- GLID$^2$E: A Gradient-Free Lightweight Fine-tune Approach for Discrete Biological Sequence Design
- HYPRL: Reinforcement Learning of Control Policies for Hyperproperties
- Learning from Demonstrations via Capability-Aware Goal Sampling
- MOOSE-Chem2: Exploring LLM Limits in Fine-Grained Scientific Hypothesis Discovery via Hierarchical Search
- Reward-Aware Proto-Representations in Reinforcement Learning
- Time Reversal Symmetry for Efficient Robotic Manipulations in Deep Reinforcement Learning