Can a Unimodal Language Agent Provide Preferences to Tune a Multimodal Vision-Language Model?

TL;DR

This paper investigates whether a text-only language model can effectively guide a multimodal vision-language model through feedback, improving its descriptions and understanding of multimodal content.

Contribution

It introduces a method for a unimodal language model to provide feedback to a vision-language model, enhancing its multimodal description capabilities.

Findings

VLM description quality improved with LLM feedback

Maximum 13% accuracy increase in multimodal tasks

Human preferences aligned with LLM feedback at 64.6%

Abstract

To explore a more scalable path for adding multimodal capabilities to existing LLMs, this paper addresses a fundamental question: Can a unimodal LLM, relying solely on text, reason about its own informational needs and provide effective feedback to optimize a multimodal model? To answer this, we propose a method that enables a language agent to give feedback to a vision-language model (VLM) to adapt text generation to the agent's preferences. Our results from different experiments affirm this hypothesis, showing that LLM preference feedback significantly enhances VLM descriptions. Using our proposed method, we find that the VLM can generate multimodal scene descriptions to help the LLM better understand multimodal context, leading to improvements of maximum 13% in absolute accuracy compared to the baseline multimodal approach. Furthermore, a human study validated our AI-driven feedback, showing a 64.6% preference alignment rate between the LLM's choices and human judgments. Extensive experiments provide insights on how and why the method works and its limitations.