Domain Adaptation with Adversarial Training and Graph Embeddings

<p dir="ltr">The success of deep neural networks (DNNs) is heavily dependent on the availability of labeled data. However, obtaining labeled data is a big challenge in many real-world problems. In such scenarios, a DNN model can leverage labeled and unlabeled data from a related doma...

Full description

Saved in:
Bibliographic Details
Main Author: Firoj Alam (14158866) (author)
Other Authors: Shafiq Joty (4576078) (author), Muhammad Imran (282621) (author)
Published: 2018
Subjects:
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1864513509783502848
author Firoj Alam (14158866)
author2 Shafiq Joty (4576078)
Muhammad Imran (282621)
author2_role author
author
author_facet Firoj Alam (14158866)
Shafiq Joty (4576078)
Muhammad Imran (282621)
author_role author
dc.creator.none.fl_str_mv Firoj Alam (14158866)
Shafiq Joty (4576078)
Muhammad Imran (282621)
dc.date.none.fl_str_mv 2018-07-01T03:00:00Z
dc.identifier.none.fl_str_mv 10.18653/v1/p18-1099
dc.relation.none.fl_str_mv https://figshare.com/articles/conference_contribution/Domain_Adaptation_with_Adversarial_Training_and_Graph_Embeddings/25919455
dc.rights.none.fl_str_mv CC BY 4.0
info:eu-repo/semantics/openAccess
dc.subject.none.fl_str_mv Information and computing sciences
Artificial intelligence
Machine learning
Language, communication and culture
Linguistics
Labeled Data
Unlabeled Data
Domain Shift
Data Distribution
Social Media Posts
Crisis Event Classification
Adversarial Learning
dc.title.none.fl_str_mv Domain Adaptation with Adversarial Training and Graph Embeddings
dc.type.none.fl_str_mv Text
Conference contribution
info:eu-repo/semantics/publishedVersion
text
conference object
description <p dir="ltr">The success of deep neural networks (DNNs) is heavily dependent on the availability of labeled data. However, obtaining labeled data is a big challenge in many real-world problems. In such scenarios, a DNN model can leverage labeled and unlabeled data from a related domain, but it has to deal with the shift in data distributions between the source and the target domains. In this paper, we study the problem of classifying social media posts during a crisis event (e.g., Earthquake). For that, we use labeled and unlabeled data from past similar events (e.g., Flood) and unlabeled data for the current event. We propose a novel model that performs adversarial learning based domain adaptation to deal with distribution drifts and graph based semi-supervised learning to leverage unlabeled data within a single unified deep learning framework. Our experiments with two real-world crisis datasets collected from Twitter demonstrate significant improvements over several baselines.</p><h2>Other Information</h2><p dir="ltr">Published in: Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)<br>License: <a href="https://creativecommons.org/licenses/by/4.0/" target="_blank">https://creativecommons.org/licenses/by/4.0/</a><br>See conference contribution on publisher's website: <a href="https://dx.doi.org/10.18653/v1/p18-1099" target="_blank">https://dx.doi.org/10.18653/v1/p18-1099</a></p>
eu_rights_str_mv openAccess
id Manara2_6275b52e27cfcd3be0ae252a8b8a292f
identifier_str_mv 10.18653/v1/p18-1099
network_acronym_str Manara2
network_name_str Manara2
oai_identifier_str oai:figshare.com:article/25919455
publishDate 2018
repository.mail.fl_str_mv
repository.name.fl_str_mv
repository_id_str
rights_invalid_str_mv CC BY 4.0
spelling Domain Adaptation with Adversarial Training and Graph EmbeddingsFiroj Alam (14158866)Shafiq Joty (4576078)Muhammad Imran (282621)Information and computing sciencesArtificial intelligenceMachine learningLanguage, communication and cultureLinguisticsLabeled DataUnlabeled DataDomain ShiftData DistributionSocial Media PostsCrisis Event ClassificationAdversarial Learning<p dir="ltr">The success of deep neural networks (DNNs) is heavily dependent on the availability of labeled data. However, obtaining labeled data is a big challenge in many real-world problems. In such scenarios, a DNN model can leverage labeled and unlabeled data from a related domain, but it has to deal with the shift in data distributions between the source and the target domains. In this paper, we study the problem of classifying social media posts during a crisis event (e.g., Earthquake). For that, we use labeled and unlabeled data from past similar events (e.g., Flood) and unlabeled data for the current event. We propose a novel model that performs adversarial learning based domain adaptation to deal with distribution drifts and graph based semi-supervised learning to leverage unlabeled data within a single unified deep learning framework. Our experiments with two real-world crisis datasets collected from Twitter demonstrate significant improvements over several baselines.</p><h2>Other Information</h2><p dir="ltr">Published in: Proceedings of the 56th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)<br>License: <a href="https://creativecommons.org/licenses/by/4.0/" target="_blank">https://creativecommons.org/licenses/by/4.0/</a><br>See conference contribution on publisher's website: <a href="https://dx.doi.org/10.18653/v1/p18-1099" target="_blank">https://dx.doi.org/10.18653/v1/p18-1099</a></p>2018-07-01T03:00:00ZTextConference contributioninfo:eu-repo/semantics/publishedVersiontextconference object10.18653/v1/p18-1099https://figshare.com/articles/conference_contribution/Domain_Adaptation_with_Adversarial_Training_and_Graph_Embeddings/25919455CC BY 4.0info:eu-repo/semantics/openAccessoai:figshare.com:article/259194552018-07-01T03:00:00Z
spellingShingle Domain Adaptation with Adversarial Training and Graph Embeddings
Firoj Alam (14158866)
Information and computing sciences
Artificial intelligence
Machine learning
Language, communication and culture
Linguistics
Labeled Data
Unlabeled Data
Domain Shift
Data Distribution
Social Media Posts
Crisis Event Classification
Adversarial Learning
status_str publishedVersion
title Domain Adaptation with Adversarial Training and Graph Embeddings
title_full Domain Adaptation with Adversarial Training and Graph Embeddings
title_fullStr Domain Adaptation with Adversarial Training and Graph Embeddings
title_full_unstemmed Domain Adaptation with Adversarial Training and Graph Embeddings
title_short Domain Adaptation with Adversarial Training and Graph Embeddings
title_sort Domain Adaptation with Adversarial Training and Graph Embeddings
topic Information and computing sciences
Artificial intelligence
Machine learning
Language, communication and culture
Linguistics
Labeled Data
Unlabeled Data
Domain Shift
Data Distribution
Social Media Posts
Crisis Event Classification
Adversarial Learning