reward model
A framework in reinforcement learning that predicts the expected rewards for given states or actions. It is typically utilized to guide decision-making processes and optimize behaviors through reinforcement signals.
- A Smooth Sea Never Made a Skilled SAILOR: Robust Imitation via Learning to Search
- Ask a Strong LLM Judge when Your Reward Model is Uncertain
- Inference-Time Reward Hacking in Large Language Models
- Inference-time Alignment in Continuous Space
- Information-Theoretic Reward Decomposition for Generalizable RLHF
- OmniGaze: Reward-inspired Generalizable Gaze Estimation in the Wild
- Online Feedback Efficient Active Target Discovery in Partially Observable Environments
- Pre-Trained Policy Discriminators are General Reward Models
- Predicting Empirical AI Research Outcomes with Language Models
- RL Tango: Reinforcing Generator and Verifier Together for Language Reasoning
- STAR: Efficient Preference-based Reinforcement Learning via Dual Regularization
- Selftok-Zero: Reinforcement Learning for Visual Generation via Discrete and Autoregressive Visual Tokens
- UI-Genie: A Self-Improving Approach for Iteratively Boosting MLLM-based Mobile GUI Agents
- Value Gradient Guidance for Flow Matching Alignment
- What Makes a Reward Model a Good Teacher? An Optimization Perspective