Unified Reinforcement and Imitation Learning for Vision-Language Models

Researchers introduced Unified Reinforcement and Imitation Learning (RIL), a training algorithm that combines reinforcement learning with adversarial imitation learning to create lightweight vision-language models (VLMs) that rival larger models. RIL enables smaller student VLMs to mimic large teacher models and improve via reinforcement signals, using an LLM-based discriminator and multiple teacher VLMs. Experiments on diverse benchmarks show RIL narrows the performance gap with state-of-the-art VLMs and sometimes surpasses them.

Vision-Language Models VLMs have achieved remarkable progress, yet their large scale often renders them impractical for resource-constrained environments. This paper introduces Unified Reinforcement and Imitation Learning RIL , a novel and efficient training algorithm designed to create powerful, lightweight VLMs. RIL distinctively combines the strengths of reinforcement learning with adversarial imitation learning. This enables smaller student VLMs not only to mimic the sophisticated text generation of large teacher models but also to systematically improve their generative capabilities through reinforcement signals. Key to our imitation framework is an LLM-based discriminator that adeptly distinguishes between student and teacher outputs, complemented by guidance from multiple large teacher VLMs to ensure diverse learning. This unified learning strategy, leveraging both reinforcement and imitation, empowers student models to achieve significant performance gains, making them competitive with leading closed-source VLMs. Extensive experiments on diverse vision-language benchmarks demonstrate that RIL significantly narrows the performance gap with state-of-the-art open- and closed-source VLMs and, in several instances, surpasses them.