@inproceedings{wang-etal-2023-scene,
title = "Scene-robust Natural Language Video Localization via Learning Domain-invariant Representations",
author = "Wang, Zehan and
Zhao, Yang and
Huang, Haifeng and
Xia, Yan and
Zhao, Zhou",
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.11",
doi = "10.18653/v1/2023.findings-acl.11",
pages = "144--160",
abstract = "Natural language video localization(NLVL) task involves the semantic matching of a text query with a moment from an untrimmed video. Previous methods primarily focus on improving performance with the assumption of independently identical data distribution while ignoring the out-of-distribution data. Therefore, these approaches often fail when handling the videos and queries in novel scenes, which is inevitable in real-world scenarios. In this paper, we, for the first time, formulate the scene-robust NLVL problem and propose a novel generalizable NLVL framework utilizing data in multiple available scenes to learn a robust model. Specifically, our model learns a group of generalizable domain-invariant representations by alignment and decomposition. First, we propose a comprehensive intra- and inter-sample distance metric for complex multi-modal feature space, and an asymmetric multi-modal alignment loss for different information densities of text and vision. Further, to alleviate the conflict between domain-invariant features for generalization and domain-specific information for reasoning, we introduce domain-specific and domain-agnostic predictors to decompose and refine the learned features by dynamically adjusting the weights of samples. Based on the original video tags, we conduct extensive experiments on three NLVL datasets with different-grained scene shifts to show the effectiveness of our proposed methods.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="wang-etal-2023-scene">
<titleInfo>
<title>Scene-robust Natural Language Video Localization via Learning Domain-invariant Representations</title>
</titleInfo>
<name type="personal">
<namePart type="given">Zehan</namePart>
<namePart type="family">Wang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yang</namePart>
<namePart type="family">Zhao</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Haifeng</namePart>
<namePart type="family">Huang</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Yan</namePart>
<namePart type="family">Xia</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Zhou</namePart>
<namePart type="family">Zhao</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>Natural language video localization(NLVL) task involves the semantic matching of a text query with a moment from an untrimmed video. Previous methods primarily focus on improving performance with the assumption of independently identical data distribution while ignoring the out-of-distribution data. Therefore, these approaches often fail when handling the videos and queries in novel scenes, which is inevitable in real-world scenarios. In this paper, we, for the first time, formulate the scene-robust NLVL problem and propose a novel generalizable NLVL framework utilizing data in multiple available scenes to learn a robust model. Specifically, our model learns a group of generalizable domain-invariant representations by alignment and decomposition. First, we propose a comprehensive intra- and inter-sample distance metric for complex multi-modal feature space, and an asymmetric multi-modal alignment loss for different information densities of text and vision. Further, to alleviate the conflict between domain-invariant features for generalization and domain-specific information for reasoning, we introduce domain-specific and domain-agnostic predictors to decompose and refine the learned features by dynamically adjusting the weights of samples. Based on the original video tags, we conduct extensive experiments on three NLVL datasets with different-grained scene shifts to show the effectiveness of our proposed methods.</abstract>
<identifier type="citekey">wang-etal-2023-scene</identifier>
<identifier type="doi">10.18653/v1/2023.findings-acl.11</identifier>
<location>
<url>https://aclanthology.org/2023.findings-acl.11</url>
</location>
<part>
<date>2023-07</date>
<extent unit="page">
<start>144</start>
<end>160</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Scene-robust Natural Language Video Localization via Learning Domain-invariant Representations
%A Wang, Zehan
%A Zhao, Yang
%A Huang, Haifeng
%A Xia, Yan
%A Zhao, Zhou
%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 wang-etal-2023-scene
%X Natural language video localization(NLVL) task involves the semantic matching of a text query with a moment from an untrimmed video. Previous methods primarily focus on improving performance with the assumption of independently identical data distribution while ignoring the out-of-distribution data. Therefore, these approaches often fail when handling the videos and queries in novel scenes, which is inevitable in real-world scenarios. In this paper, we, for the first time, formulate the scene-robust NLVL problem and propose a novel generalizable NLVL framework utilizing data in multiple available scenes to learn a robust model. Specifically, our model learns a group of generalizable domain-invariant representations by alignment and decomposition. First, we propose a comprehensive intra- and inter-sample distance metric for complex multi-modal feature space, and an asymmetric multi-modal alignment loss for different information densities of text and vision. Further, to alleviate the conflict between domain-invariant features for generalization and domain-specific information for reasoning, we introduce domain-specific and domain-agnostic predictors to decompose and refine the learned features by dynamically adjusting the weights of samples. Based on the original video tags, we conduct extensive experiments on three NLVL datasets with different-grained scene shifts to show the effectiveness of our proposed methods.
%R 10.18653/v1/2023.findings-acl.11
%U https://aclanthology.org/2023.findings-acl.11
%U https://doi.org/10.18653/v1/2023.findings-acl.11
%P 144-160
Markdown (Informal)
[Scene-robust Natural Language Video Localization via Learning Domain-invariant Representations](https://aclanthology.org/2023.findings-acl.11) (Wang et al., Findings 2023)
ACL