Search papers, labs, and topics across Lattice.
Reward-Zero is introduced as an implicit reward mechanism that uses language embeddings to provide dense, semantically grounded progress signals for RL agents. It compares the embedding of a task description with embeddings derived from the agent's interaction experience, generating a continuous "sense-of-completion" reward signal. Experiments show that integrating Reward-Zero into PPO accelerates exploration, stabilizes training, and enhances generalization, leading to faster convergence and higher success rates compared to hand-designed reward shaping, especially in complex tasks.
Forget hand-engineered reward functions: Reward-Zero uses language embeddings to give RL agents an intrinsic "sense of completion," dramatically improving sample efficiency and generalization.
We introduce Reward-Zero, a general-purpose implicit reward mechanism that transforms natural-language task descriptions into dense, semantically grounded progress signals for reinforcement learning (RL). Reward-Zero serves as a simple yet sophisticated universal reward function that leverages language embeddings for efficient RL training. By comparing the embedding of a task specification with embeddings derived from an agent's interaction experience, Reward-Zero produces a continuous, semantically aligned sense-of-completion signal. This reward supplements sparse or delayed environmental feedback without requiring task-specific engineering. When integrated into standard RL frameworks, it accelerates exploration, stabilizes training, and enhances generalization across diverse tasks. Empirically, agents trained with Reward-Zero converge faster and achieve higher final success rates than conventional methods such as PPO with common reward-shaping baselines, successfully solving tasks that hand-designed rewards could not in some complex tasks. In addition, we develop a mini benchmark for the evaluation of completion sense during task execution via language embeddings. These results highlight the promise of language-driven implicit reward functions as a practical path toward more sample-efficient, generalizable, and scalable RL for embodied agents. Code will be released after peer review.
