MAGID: An Automated Pipeline for Generating Synthetic Multi-modal Datasets

Authors:
  • HAHossein Aboutalebi
    Hossein Aboutalebi
  • HSHwanjun Song
    Hwanjun Song
  • YXYusheng Xie
    Yusheng Xie
  • AGArshit Gupta
    Arshit Gupta
  • JSJustin Sun
    Justin Sun
  • HSHang Su
    Hang Su
  • ISIgor Shalyminov
    Igor Shalyminov
  • NPNikolaos Pappas
    Nikolaos Pappas
  • SSSiffi Singh
    Siffi Singh
  • SMSaab Mansour
    Saab Mansour

Abstract
Development of multimodal interactive systems is hindered by the lack of rich, multimodal (text, images) conversational data, which is needed in large quantities for LLMs. Previous approaches augment textual dialogues with retrieved images, posing privacy, diversity, and quality constraints. In this work, we introduce Multimodal Augmented Generative Images Dialogues (MAGID), a framework to augment text-only dialogues with diverse and high-quality images. Subsequently, a diffusion model is applied to craft corresponding images, ensuring alignment with the identified text. Finally, MAGID incorporates an innovative feedback loop between an image description generation module (textual LLM) and image quality modules (addressing aesthetics, image-text matching, and safety), that work in tandem to generate high-quality and multi-modal dialogues. We compare MAGID to other SOTA baselines on three dialogue datasets, using automated and human evaluation. Our results show that MAGID is comparable to or better than baselines, with significant improvements in human evaluation, especially against retrieval baselines where the image database is small.