A Multimodal Recaptioning Framework to Account for Perceptual Diversity Across Languages in Vision-Language Modeling
This addresses bias in multilingual vision-language models for users across languages and cultures, though it is incremental as it builds on existing methods.
The paper tackles perceptual bias in vision-language models by proposing a multimodal recaptioning framework that uses native speaker data and multimodal LLM reasoning to augment captions for target languages, resulting in improvements of up to +3.5 mean recall and +4.4 on native vs. translation errors in German and Japanese text-image retrieval.
When captioning an image, people describe objects in diverse ways, such as by using different terms and/or including details that are perceptually noteworthy to them. Descriptions can be especially unique across languages and cultures. Modern vision-language models (VLMs) gain understanding of images with text in different languages often through training on machine translations of English captions. However, this process relies on input content written from the perception of English speakers, leading to a perceptual bias. In this work, we outline a framework to address this bias. We specifically use a small amount of native speaker data, nearest-neighbor example guidance, and multimodal LLM reasoning to augment captions to better reflect descriptions in a target language. When adding the resulting rewrites to multilingual CLIP finetuning, we improve on German and Japanese text-image retrieval case studies (up to +3.5 mean recall, +4.4 on native vs. translation errors). We also propose a mechanism to build understanding of object description variation across languages, and offer insights into cross-dataset and cross-language generalization.