@inproceedings{rosch-etal-2024-enhancing,
    title = "Enhancing Conceptual Understanding in Multimodal Contrastive Learning through Hard Negative Samples",
    author = {R{\"o}sch, Philipp J.  and
      Oswald, Norbert  and
      Geierhos, Michaela  and
      Libovick{\'y}, Jind{\v{r}}ich},
    editor = "Gu, Jing  and
      Fu, Tsu-Jui (Ray)  and
      Hudson, Drew  and
      Celikyilmaz, Asli  and
      Wang, William",
    booktitle = "Proceedings of the 3rd Workshop on Advances in Language and Vision Research (ALVR)",
    month = aug,
    year = "2024",
    address = "Bangkok, Thailand",
    publisher = "Association for Computational Linguistics",
    url = "https://preview.aclanthology.org/jlcl-multiple-ingestion/2024.alvr-1.9/",
    doi = "10.18653/v1/2024.alvr-1.9",
    pages = "102--115",
    abstract = "Current vision-language models leveraging contrastive learning often face limitations in developing fine-grained conceptual understanding. This is due to random negative samples during pretraining, causing almost exclusively very dissimilar concepts to be compared in the loss function. Consequently, the models struggle with fine-grained semantic differences. To address this problem, we introduce a novel pretraining method incorporating synthetic hard negative text examples. The hard negatives replace terms corresponding to visual concepts, leading to a more fine-grained visual and textual concept alignment. Further, we introduce InpaintCOCO, a new challenging dataset for assessing the fine-grained alignment of colors, objects, and sizes in vision-language models. We created the dataset using generative inpainting from COCO images by changing the visual concepts so that the images no longer match their original captions. Our results show significant improvements in fine-grained concept understanding across various vision-language datasets, including our InpaintCOCO dataset."
}