@inproceedings{khandelwal-etal-2023-large,
title = "Large Scale Generative Multimodal Attribute Extraction for {E}-commerce Attributes",
author = "Khandelwal, Anant and
Mittal, Happy and
Kulkarni, Shreyas and
Gupta, Deepak",
editor = "Sitaram, Sunayana and
Beigman Klebanov, Beata and
Williams, Jason D",
booktitle = "Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 5: Industry Track)",
month = jul,
year = "2023",
address = "Toronto, Canada",
publisher = "Association for Computational Linguistics",
url = "https://aclanthology.org/2023.acl-industry.29",
doi = "10.18653/v1/2023.acl-industry.29",
pages = "305--312",
abstract = "E-commerce websites (e.g. Amazon, Alibaba) have a plethora of structured and unstructured information (text and images) present on the product pages. Sellers often don{'}t label or mislabel values of the attributes (e.g. color, size etc.) for their products. Automatically identifying these attribute values from an eCommerce product page that contains both text and images is a challenging task, especially when the attribute value is not explicitly mentioned in the catalog. In this paper, we present a scalable solution for this problem where we pose attribute extraction problem as a question-answering task, which we solve using MXT, that consists of three key components: (i) MAG (Multimodal Adaptation Gate), (ii) Xception network, and (iii) T5 encoder-decoder. Our system consists of a generative model that generates attribute-values for a given product by using both textual and visual characteristics (e.g. images) of the product. We show that our system is capable of handling zero-shot attribute prediction (when attribute value is not seen in training data) and value-absent prediction (when attribute value is not mentioned in the text) which are missing in traditional classification-based and NER-based models respectively. We have trained our models using distant supervision, removing dependency on human labeling, thus making them practical for real-world applications. With this framework, we are able to train a single model for 1000s of (product-type, attribute) pairs, thus reducing the overhead of training and maintaining separate models. Extensive experiments on two real world datasets (total 57 attributes) show that our framework improves the absolute recall@90P by 10.16{\%} and 6.9 from the existing state of the art models. In a popular e-commerce store, we have productionized our models that cater to 12K (product-type, attribute) pairs, and have extracted 150MM attribute values.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="khandelwal-etal-2023-large">
<titleInfo>
<title>Large Scale Generative Multimodal Attribute Extraction for E-commerce Attributes</title>
</titleInfo>
<name type="personal">
<namePart type="given">Anant</namePart>
<namePart type="family">Khandelwal</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Happy</namePart>
<namePart type="family">Mittal</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Shreyas</namePart>
<namePart type="family">Kulkarni</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Deepak</namePart>
<namePart type="family">Gupta</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>Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 5: Industry Track)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Sunayana</namePart>
<namePart type="family">Sitaram</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Beata</namePart>
<namePart type="family">Beigman Klebanov</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Jason</namePart>
<namePart type="given">D</namePart>
<namePart type="family">Williams</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>E-commerce websites (e.g. Amazon, Alibaba) have a plethora of structured and unstructured information (text and images) present on the product pages. Sellers often don’t label or mislabel values of the attributes (e.g. color, size etc.) for their products. Automatically identifying these attribute values from an eCommerce product page that contains both text and images is a challenging task, especially when the attribute value is not explicitly mentioned in the catalog. In this paper, we present a scalable solution for this problem where we pose attribute extraction problem as a question-answering task, which we solve using MXT, that consists of three key components: (i) MAG (Multimodal Adaptation Gate), (ii) Xception network, and (iii) T5 encoder-decoder. Our system consists of a generative model that generates attribute-values for a given product by using both textual and visual characteristics (e.g. images) of the product. We show that our system is capable of handling zero-shot attribute prediction (when attribute value is not seen in training data) and value-absent prediction (when attribute value is not mentioned in the text) which are missing in traditional classification-based and NER-based models respectively. We have trained our models using distant supervision, removing dependency on human labeling, thus making them practical for real-world applications. With this framework, we are able to train a single model for 1000s of (product-type, attribute) pairs, thus reducing the overhead of training and maintaining separate models. Extensive experiments on two real world datasets (total 57 attributes) show that our framework improves the absolute recall@90P by 10.16% and 6.9 from the existing state of the art models. In a popular e-commerce store, we have productionized our models that cater to 12K (product-type, attribute) pairs, and have extracted 150MM attribute values.</abstract>
<identifier type="citekey">khandelwal-etal-2023-large</identifier>
<identifier type="doi">10.18653/v1/2023.acl-industry.29</identifier>
<location>
<url>https://aclanthology.org/2023.acl-industry.29</url>
</location>
<part>
<date>2023-07</date>
<extent unit="page">
<start>305</start>
<end>312</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Large Scale Generative Multimodal Attribute Extraction for E-commerce Attributes
%A Khandelwal, Anant
%A Mittal, Happy
%A Kulkarni, Shreyas
%A Gupta, Deepak
%Y Sitaram, Sunayana
%Y Beigman Klebanov, Beata
%Y Williams, Jason D.
%S Proceedings of the 61st Annual Meeting of the Association for Computational Linguistics (Volume 5: Industry Track)
%D 2023
%8 July
%I Association for Computational Linguistics
%C Toronto, Canada
%F khandelwal-etal-2023-large
%X E-commerce websites (e.g. Amazon, Alibaba) have a plethora of structured and unstructured information (text and images) present on the product pages. Sellers often don’t label or mislabel values of the attributes (e.g. color, size etc.) for their products. Automatically identifying these attribute values from an eCommerce product page that contains both text and images is a challenging task, especially when the attribute value is not explicitly mentioned in the catalog. In this paper, we present a scalable solution for this problem where we pose attribute extraction problem as a question-answering task, which we solve using MXT, that consists of three key components: (i) MAG (Multimodal Adaptation Gate), (ii) Xception network, and (iii) T5 encoder-decoder. Our system consists of a generative model that generates attribute-values for a given product by using both textual and visual characteristics (e.g. images) of the product. We show that our system is capable of handling zero-shot attribute prediction (when attribute value is not seen in training data) and value-absent prediction (when attribute value is not mentioned in the text) which are missing in traditional classification-based and NER-based models respectively. We have trained our models using distant supervision, removing dependency on human labeling, thus making them practical for real-world applications. With this framework, we are able to train a single model for 1000s of (product-type, attribute) pairs, thus reducing the overhead of training and maintaining separate models. Extensive experiments on two real world datasets (total 57 attributes) show that our framework improves the absolute recall@90P by 10.16% and 6.9 from the existing state of the art models. In a popular e-commerce store, we have productionized our models that cater to 12K (product-type, attribute) pairs, and have extracted 150MM attribute values.
%R 10.18653/v1/2023.acl-industry.29
%U https://aclanthology.org/2023.acl-industry.29
%U https://doi.org/10.18653/v1/2023.acl-industry.29
%P 305-312
Markdown (Informal)
[Large Scale Generative Multimodal Attribute Extraction for E-commerce Attributes](https://aclanthology.org/2023.acl-industry.29) (Khandelwal et al., ACL 2023)
ACL