Vision-Language Models are Zero-Shot Reward Models for Reinforcement Learning.
Juan RocamondeVictoriano MontesinosElvis NavaEthan PerezDavid LindnerPublished in: ICLR (2024)
Keyphrases
- language model
- reinforcement learning
- probabilistic model
- language modelling
- language modeling
- statistical language models
- speech recognition
- relevance model
- smoothing methods
- document retrieval
- translation model
- test collection
- statistical models
- context sensitive
- ad hoc information retrieval
- retrieval model
- optimal policy
- generative model
- statistical language modeling
- language models for information retrieval