@inproceedings{jiang-etal-2024-comclip,
title = "{C}om{CLIP}: Training-Free Compositional Image and Text Matching",
author = "Jiang, Kenan and
He, Xuehai and
Xu, Ruize and
Wang, Xin",
editor = "Duh, Kevin and
Gomez, Helena and
Bethard, Steven",
booktitle = "Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)",
month = jun,
year = "2024",
address = "Mexico City, Mexico",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2024.naacl-long.370/",
doi = "10.18653/v1/2024.naacl-long.370",
pages = "6639--6659",
abstract = "Contrastive Language-Image Pretraining (CLIP) has demonstrated great zero-shot performance for matching images and text. However, it is still challenging to adapt vision-language pretrained models like CLIP to compositional image and text matching {---} a more challenging image and text matching task requiring the model`s understanding of compositional word concepts and visual components. Towards better compositional generalization in zero-shot image and text matching, in this paper, we study the problem from a causal perspective: the erroneous semantics of individual entities are essentially confounders that cause the matching failure. Therefore, we propose a novel training-free compositional CLIP model (ComCLIP). ComCLIP disentangles input images into subjects, objects, and action subimages and composes CLIP`s vision encoder and text encoder to perform evolving matching over compositional text embedding and subimage embeddings. In this way, ComCLIP can mitigate spurious correlations introduced by the pretrained CLIP models and dynamically evaluate the importance of each component. Experiments on four compositional image-text matching datasets: Winoground, VL-checklist, SVO, and ComVG, and two general image-text retrieval datasets: Flick30K, and MSCOCO demonstrate the effectiveness of our plug-and-play method, which boosts the zero-shot inference ability of CLIP, SLIP, and BLIP2 even without further training or fine-tuning. Our codes can be found at https://github.com/eric-ai-lab/ComCLIP."
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="jiang-etal-2024-comclip">
<titleInfo>
<title>ComCLIP: Training-Free Compositional Image and Text Matching</title>
</titleInfo>
<name type="personal">
<namePart type="given">Kenan</namePart>
<namePart type="family">Jiang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xuehai</namePart>
<namePart type="family">He</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ruize</namePart>
<namePart type="family">Xu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Xin</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2024-06</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Kevin</namePart>
<namePart type="family">Duh</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Helena</namePart>
<namePart type="family">Gomez</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Steven</namePart>
<namePart type="family">Bethard</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Mexico City, Mexico</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>Contrastive Language-Image Pretraining (CLIP) has demonstrated great zero-shot performance for matching images and text. However, it is still challenging to adapt vision-language pretrained models like CLIP to compositional image and text matching — a more challenging image and text matching task requiring the model‘s understanding of compositional word concepts and visual components. Towards better compositional generalization in zero-shot image and text matching, in this paper, we study the problem from a causal perspective: the erroneous semantics of individual entities are essentially confounders that cause the matching failure. Therefore, we propose a novel training-free compositional CLIP model (ComCLIP). ComCLIP disentangles input images into subjects, objects, and action subimages and composes CLIP‘s vision encoder and text encoder to perform evolving matching over compositional text embedding and subimage embeddings. In this way, ComCLIP can mitigate spurious correlations introduced by the pretrained CLIP models and dynamically evaluate the importance of each component. Experiments on four compositional image-text matching datasets: Winoground, VL-checklist, SVO, and ComVG, and two general image-text retrieval datasets: Flick30K, and MSCOCO demonstrate the effectiveness of our plug-and-play method, which boosts the zero-shot inference ability of CLIP, SLIP, and BLIP2 even without further training or fine-tuning. Our codes can be found at https://github.com/eric-ai-lab/ComCLIP.</abstract>
<identifier type="citekey">jiang-etal-2024-comclip</identifier>
<identifier type="doi">10.18653/v1/2024.naacl-long.370</identifier>
<location>
<url>https://aclanthology.org/2024.naacl-long.370/</url>
</location>
<part>
<date>2024-06</date>
<extent unit="page">
<start>6639</start>
<end>6659</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T ComCLIP: Training-Free Compositional Image and Text Matching
%A Jiang, Kenan
%A He, Xuehai
%A Xu, Ruize
%A Wang, Xin
%Y Duh, Kevin
%Y Gomez, Helena
%Y Bethard, Steven
%S Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers)
%D 2024
%8 June
%I Association for Computational Linguistics
%C Mexico City, Mexico
%F jiang-etal-2024-comclip
%X Contrastive Language-Image Pretraining (CLIP) has demonstrated great zero-shot performance for matching images and text. However, it is still challenging to adapt vision-language pretrained models like CLIP to compositional image and text matching — a more challenging image and text matching task requiring the model‘s understanding of compositional word concepts and visual components. Towards better compositional generalization in zero-shot image and text matching, in this paper, we study the problem from a causal perspective: the erroneous semantics of individual entities are essentially confounders that cause the matching failure. Therefore, we propose a novel training-free compositional CLIP model (ComCLIP). ComCLIP disentangles input images into subjects, objects, and action subimages and composes CLIP‘s vision encoder and text encoder to perform evolving matching over compositional text embedding and subimage embeddings. In this way, ComCLIP can mitigate spurious correlations introduced by the pretrained CLIP models and dynamically evaluate the importance of each component. Experiments on four compositional image-text matching datasets: Winoground, VL-checklist, SVO, and ComVG, and two general image-text retrieval datasets: Flick30K, and MSCOCO demonstrate the effectiveness of our plug-and-play method, which boosts the zero-shot inference ability of CLIP, SLIP, and BLIP2 even without further training or fine-tuning. Our codes can be found at https://github.com/eric-ai-lab/ComCLIP.
%R 10.18653/v1/2024.naacl-long.370
%U https://aclanthology.org/2024.naacl-long.370/
%U https://doi.org/10.18653/v1/2024.naacl-long.370
%P 6639-6659
Markdown (Informal)
[ComCLIP: Training-Free Compositional Image and Text Matching](https://aclanthology.org/2024.naacl-long.370/) (Jiang et al., NAACL 2024)
ACL
- Kenan Jiang, Xuehai He, Ruize Xu, and Xin Wang. 2024. ComCLIP: Training-Free Compositional Image and Text Matching. In Proceedings of the 2024 Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (Volume 1: Long Papers), pages 6639–6659, Mexico City, Mexico. Association for Computational Linguistics.