@inproceedings{chakraborty-majumder-2021-deep,
title = "Deep Embedding of Conversation Segments",
author = "Chakraborty, Abir and
Majumder, Anirban",
editor = "Bandyopadhyay, Sivaji and
Devi, Sobha Lalitha and
Bhattacharyya, Pushpak",
booktitle = "Proceedings of the 18th International Conference on Natural Language Processing (ICON)",
month = dec,
year = "2021",
address = "National Institute of Technology Silchar, Silchar, India",
publisher = "NLP Association of India (NLPAI)",
url = "https://aclanthology.org/2021.icon-main.68",
pages = "555--563",
abstract = "We introduce a novel conversation embedding by extending Bidirectional Encoder Representations from Transformers (BERT) framework. Specifically, information related to {``}turn{''} and {``}role{''} that are unique to conversations are augmented to the word tokens and the next sentence prediction task predicts a segment of a conversation possibly spanning across multiple roles and turns. It is observed that the addition of role and turn substantially increases the next sentence prediction accuracy. Conversation embeddings obtained in this fashion are applied to (a) conversation clustering, (b) conversation classification and (c) as a context for automated conversation generation on new datasets (unseen by the pre-training model). We found that clustering accuracy is greatly improved if embeddings are used as features as opposed to conventional tf-idf based features that do not take role or turn information into account. On classification task, a fine-tuned model on conversation embedding achieves accuracy comparable to an optimized linear SVM model on tf-idf based features. Finally, we present a way of capturing variable length context in sequence-to-sequence models by utilizing this conversation embedding and show that BLEU score improves over a vanilla sequence to sequence model without context.",
}
<?xml version="1.0" encoding="UTF-8"?>
<modsCollection xmlns="http://www.loc.gov/mods/v3">
<mods ID="chakraborty-majumder-2021-deep">
<titleInfo>
<title>Deep Embedding of Conversation Segments</title>
</titleInfo>
<name type="personal">
<namePart type="given">Abir</namePart>
<namePart type="family">Chakraborty</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Anirban</namePart>
<namePart type="family">Majumder</namePart>
<role>
<roleTerm authority="marcrelator" type="text">author</roleTerm>
</role>
</name>
<originInfo>
<dateIssued>2021-12</dateIssued>
</originInfo>
<typeOfResource>text</typeOfResource>
<relatedItem type="host">
<titleInfo>
<title>Proceedings of the 18th International Conference on Natural Language Processing (ICON)</title>
</titleInfo>
<name type="personal">
<namePart type="given">Sivaji</namePart>
<namePart type="family">Bandyopadhyay</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Sobha</namePart>
<namePart type="given">Lalitha</namePart>
<namePart type="family">Devi</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<name type="personal">
<namePart type="given">Pushpak</namePart>
<namePart type="family">Bhattacharyya</namePart>
<role>
<roleTerm authority="marcrelator" type="text">editor</roleTerm>
</role>
</name>
<originInfo>
<publisher>NLP Association of India (NLPAI)</publisher>
<place>
<placeTerm type="text">National Institute of Technology Silchar, Silchar, India</placeTerm>
</place>
</originInfo>
<genre authority="marcgt">conference publication</genre>
</relatedItem>
<abstract>We introduce a novel conversation embedding by extending Bidirectional Encoder Representations from Transformers (BERT) framework. Specifically, information related to “turn” and “role” that are unique to conversations are augmented to the word tokens and the next sentence prediction task predicts a segment of a conversation possibly spanning across multiple roles and turns. It is observed that the addition of role and turn substantially increases the next sentence prediction accuracy. Conversation embeddings obtained in this fashion are applied to (a) conversation clustering, (b) conversation classification and (c) as a context for automated conversation generation on new datasets (unseen by the pre-training model). We found that clustering accuracy is greatly improved if embeddings are used as features as opposed to conventional tf-idf based features that do not take role or turn information into account. On classification task, a fine-tuned model on conversation embedding achieves accuracy comparable to an optimized linear SVM model on tf-idf based features. Finally, we present a way of capturing variable length context in sequence-to-sequence models by utilizing this conversation embedding and show that BLEU score improves over a vanilla sequence to sequence model without context.</abstract>
<identifier type="citekey">chakraborty-majumder-2021-deep</identifier>
<location>
<url>https://aclanthology.org/2021.icon-main.68</url>
</location>
<part>
<date>2021-12</date>
<extent unit="page">
<start>555</start>
<end>563</end>
</extent>
</part>
</mods>
</modsCollection>
%0 Conference Proceedings
%T Deep Embedding of Conversation Segments
%A Chakraborty, Abir
%A Majumder, Anirban
%Y Bandyopadhyay, Sivaji
%Y Devi, Sobha Lalitha
%Y Bhattacharyya, Pushpak
%S Proceedings of the 18th International Conference on Natural Language Processing (ICON)
%D 2021
%8 December
%I NLP Association of India (NLPAI)
%C National Institute of Technology Silchar, Silchar, India
%F chakraborty-majumder-2021-deep
%X We introduce a novel conversation embedding by extending Bidirectional Encoder Representations from Transformers (BERT) framework. Specifically, information related to “turn” and “role” that are unique to conversations are augmented to the word tokens and the next sentence prediction task predicts a segment of a conversation possibly spanning across multiple roles and turns. It is observed that the addition of role and turn substantially increases the next sentence prediction accuracy. Conversation embeddings obtained in this fashion are applied to (a) conversation clustering, (b) conversation classification and (c) as a context for automated conversation generation on new datasets (unseen by the pre-training model). We found that clustering accuracy is greatly improved if embeddings are used as features as opposed to conventional tf-idf based features that do not take role or turn information into account. On classification task, a fine-tuned model on conversation embedding achieves accuracy comparable to an optimized linear SVM model on tf-idf based features. Finally, we present a way of capturing variable length context in sequence-to-sequence models by utilizing this conversation embedding and show that BLEU score improves over a vanilla sequence to sequence model without context.
%U https://aclanthology.org/2021.icon-main.68
%P 555-563
Markdown (Informal)
[Deep Embedding of Conversation Segments](https://aclanthology.org/2021.icon-main.68) (Chakraborty & Majumder, ICON 2021)
ACL
- Abir Chakraborty and Anirban Majumder. 2021. Deep Embedding of Conversation Segments. In Proceedings of the 18th International Conference on Natural Language Processing (ICON), pages 555–563, National Institute of Technology Silchar, Silchar, India. NLP Association of India (NLPAI).