@inproceedings{chen-etal-2023-altclip,
title = "{A}lt{CLIP}: Altering the Language Encoder in {CLIP} for Extended Language Capabilities",
author = "Chen, Zhongzhi and
Liu, Guang and
Zhang, Bo-Wen and
Yang, Qinghong and
Wu, Ledell",
editor = "Rogers, Anna and
Boyd-Graber, Jordan and
Okazaki, Naoaki",
booktitle = "Findings of the Association for Computational Linguistics: ACL 2023",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.findings-acl.552",
doi = "10.18653/v1/2023.findings-acl.552",
pages = "8666--8682",
abstract = "CLIP (Contrastive Language{--}Image Pretraining) is an English multimodal representation model learned from a massive amount of English text-image pairs and has achieved great success in various downstream tasks, including image classification, text-to-image retrieval, and image generation. When extending CLIP to other languages, the major problem is the lack of good-quality text-image pairs. In this work, we present AltCLIP, a simple and low-resource method to build a strong multilingual multimodal representation model. Instead of training a model from scratch on multilingual text-image pairs, we take the original CLIP model trained on English text-image pairs and alter its text encoder with a pre-trained multilingual text encoder (XLM-R). We then align text and image representations by a two-stage training schema consisting of teacher learning and contrastive learning. Our method utilizes the existence of rich parallel text data and pre-trained multilingual language models. We present extensive experimental evaluations to demonstrate the effectiveness of our proposed method. Our model sets new state-of-the-art zero-shot performances on a wide range of tasks in multilingual multimodal benchmarks, including ImageNet-CN/IT/JA/KO serials, Flicker30k-CN, COCO-CN, Multi30k, and XTD. Further, our model outperforms the original CLIP model on zero-shot cross-modal retrieval, Image Classification in the Wild (ICinW) tasks, and CLIP Benchmark. We plan to open-source our code, pre-trained model weights, and evaluation toolkits of multilingual multimodal tasks, to facilitate research on multilingual multimodal representation learning.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="chen-etal-2023-altclip">
<titleInfo>
<title>AltCLIP: Altering the Language Encoder in CLIP for Extended Language Capabilities</title>
</titleInfo>
<name type="personal">
<namePart type="given">Zhongzhi</namePart>
<namePart type="family">Chen</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Guang</namePart>
<namePart type="family">Liu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Bo-Wen</namePart>
<namePart type="family">Zhang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Qinghong</namePart>
<namePart type="family">Yang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Ledell</namePart>
<namePart type="family">Wu</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2023-07</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Findings of the Association for Computational Linguistics: ACL 2023</title>
</titleInfo>
<name type="personal">
<namePart type="given">Anna</namePart>
<namePart type="family">Rogers</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jordan</namePart>
<namePart type="family">Boyd-Graber</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Naoaki</namePart>
<namePart type="family">Okazaki</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>Association for Computational Linguistics</publisher>
<place>
<placeTerm type="text">Toronto, Canada</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>CLIP (Contrastive Language–Image Pretraining) is an English multimodal representation model learned from a massive amount of English text-image pairs and has achieved great success in various downstream tasks, including image classification, text-to-image retrieval, and image generation. When extending CLIP to other languages, the major problem is the lack of good-quality text-image pairs. In this work, we present AltCLIP, a simple and low-resource method to build a strong multilingual multimodal representation model. Instead of training a model from scratch on multilingual text-image pairs, we take the original CLIP model trained on English text-image pairs and alter its text encoder with a pre-trained multilingual text encoder (XLM-R). We then align text and image representations by a two-stage training schema consisting of teacher learning and contrastive learning. Our method utilizes the existence of rich parallel text data and pre-trained multilingual language models. We present extensive experimental evaluations to demonstrate the effectiveness of our proposed method. Our model sets new state-of-the-art zero-shot performances on a wide range of tasks in multilingual multimodal benchmarks, including ImageNet-CN/IT/JA/KO serials, Flicker30k-CN, COCO-CN, Multi30k, and XTD. Further, our model outperforms the original CLIP model on zero-shot cross-modal retrieval, Image Classification in the Wild (ICinW) tasks, and CLIP Benchmark. We plan to open-source our code, pre-trained model weights, and evaluation toolkits of multilingual multimodal tasks, to facilitate research on multilingual multimodal representation learning.</abstract>
<identifier type="citekey">chen-etal-2023-altclip</identifier>
<identifier type="doi">10.18653/v1/2023.findings-acl.552</identifier>
<location>
<url>https://aclanthology.org/2023.findings-acl.552</url>
</location>
<part>
<date>2023-07</date>
<extent unit="page">
<start>8666</start>
<end>8682</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T AltCLIP: Altering the Language Encoder in CLIP for Extended Language Capabilities
%A Chen, Zhongzhi
%A Liu, Guang
%A Zhang, Bo-Wen
%A Yang, Qinghong
%A Wu, Ledell
%Y Rogers, Anna
%Y Boyd-Graber, Jordan
%Y Okazaki, Naoaki
%S Findings of the Association for Computational Linguistics: ACL 2023
%D 2023
%8 July
%I Association for Computational Linguistics
%C Toronto, Canada
%F chen-etal-2023-altclip
%X CLIP (Contrastive Language–Image Pretraining) is an English multimodal representation model learned from a massive amount of English text-image pairs and has achieved great success in various downstream tasks, including image classification, text-to-image retrieval, and image generation. When extending CLIP to other languages, the major problem is the lack of good-quality text-image pairs. In this work, we present AltCLIP, a simple and low-resource method to build a strong multilingual multimodal representation model. Instead of training a model from scratch on multilingual text-image pairs, we take the original CLIP model trained on English text-image pairs and alter its text encoder with a pre-trained multilingual text encoder (XLM-R). We then align text and image representations by a two-stage training schema consisting of teacher learning and contrastive learning. Our method utilizes the existence of rich parallel text data and pre-trained multilingual language models. We present extensive experimental evaluations to demonstrate the effectiveness of our proposed method. Our model sets new state-of-the-art zero-shot performances on a wide range of tasks in multilingual multimodal benchmarks, including ImageNet-CN/IT/JA/KO serials, Flicker30k-CN, COCO-CN, Multi30k, and XTD. Further, our model outperforms the original CLIP model on zero-shot cross-modal retrieval, Image Classification in the Wild (ICinW) tasks, and CLIP Benchmark. We plan to open-source our code, pre-trained model weights, and evaluation toolkits of multilingual multimodal tasks, to facilitate research on multilingual multimodal representation learning.
%R 10.18653/v1/2023.findings-acl.552
%U https://aclanthology.org/2023.findings-acl.552
%U https://doi.org/10.18653/v1/2023.findings-acl.552
%P 8666-8682
Markdown (Informal)
[AltCLIP: Altering the Language Encoder in CLIP for Extended Language Capabilities](https://aclanthology.org/2023.findings-acl.552) (Chen et al., Findings 2023)
ACL