Towards automated customer support

Momchil Hardalov, Ivan Koychev, Preslav Nakov

Research output: Chapter in Book/Report/Conference proceedingConference contribution

2 Citations (Scopus)

Abstract

Recent years have seen growing interest in conversational agents, such as chatbots, which are a very good fit for automated customer support because the domain in which they need to operate is narrow. This interest was in part inspired by recent advances in neural machine translation, esp. the rise of sequence-to-sequence (seq2seq) and attention-based models such as the Transformer, which have been applied to various other tasks and have opened new research directions in question answering, chatbots, and conversational systems. Still, in many cases, it might be feasible and even preferable to use simple information retrieval techniques. Thus, here we compare three different models: (i) a retrieval model, (ii) a sequence-to-sequence model with attention, and (iii) Transformer. Our experiments with the Twitter Customer Support Dataset, which contains over two million posts from customer support services of twenty major brands, show that the seq2seq model outperforms the other two in terms of semantics and word overlap.

Original languageEnglish
Title of host publicationArtificial Intelligence
Subtitle of host publicationMethodology, Systems, and Applications - 18th International Conference, AIMSA 2018, Proceedings
EditorsJosef van Genabith, Gennady Agre, Thierry Declerck
PublisherSpringer Verlag
Pages48-59
Number of pages12
ISBN (Print)9783319993430
DOIs
Publication statusPublished - 1 Jan 2018
Event18th International Conference on Artificial Intelligence: Methodology, Systems, and Applications, AIMSA 2018 - Varna, Bulgaria
Duration: 12 Sep 201814 Sep 2018

Publication series

NameLecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume11089 LNAI
ISSN (Print)0302-9743
ISSN (Electronic)1611-3349

Other

Other18th International Conference on Artificial Intelligence: Methodology, Systems, and Applications, AIMSA 2018
CountryBulgaria
CityVarna
Period12/9/1814/9/18

Fingerprint

Customers
Transformer
Machine Translation
Model
Question Answering
Information retrieval
Information Retrieval
Overlap
Retrieval
Semantics
Experiment
Experiments

Keywords

  • Chatbots
  • Conversational agents
  • Customer support
  • IR
  • Seq2seq
  • Transformer

ASJC Scopus subject areas

  • Theoretical Computer Science
  • Computer Science(all)

Cite this

Hardalov, M., Koychev, I., & Nakov, P. (2018). Towards automated customer support. In J. van Genabith, G. Agre, & T. Declerck (Eds.), Artificial Intelligence: Methodology, Systems, and Applications - 18th International Conference, AIMSA 2018, Proceedings (pp. 48-59). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11089 LNAI). Springer Verlag. https://doi.org/10.1007/978-3-319-99344-7_5

Towards automated customer support. / Hardalov, Momchil; Koychev, Ivan; Nakov, Preslav.

Artificial Intelligence: Methodology, Systems, and Applications - 18th International Conference, AIMSA 2018, Proceedings. ed. / Josef van Genabith; Gennady Agre; Thierry Declerck. Springer Verlag, 2018. p. 48-59 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 11089 LNAI).

Research output: Chapter in Book/Report/Conference proceedingConference contribution

Hardalov, M, Koychev, I & Nakov, P 2018, Towards automated customer support. in J van Genabith, G Agre & T Declerck (eds), Artificial Intelligence: Methodology, Systems, and Applications - 18th International Conference, AIMSA 2018, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 11089 LNAI, Springer Verlag, pp. 48-59, 18th International Conference on Artificial Intelligence: Methodology, Systems, and Applications, AIMSA 2018, Varna, Bulgaria, 12/9/18. https://doi.org/10.1007/978-3-319-99344-7_5
Hardalov M, Koychev I, Nakov P. Towards automated customer support. In van Genabith J, Agre G, Declerck T, editors, Artificial Intelligence: Methodology, Systems, and Applications - 18th International Conference, AIMSA 2018, Proceedings. Springer Verlag. 2018. p. 48-59. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). https://doi.org/10.1007/978-3-319-99344-7_5
Hardalov, Momchil ; Koychev, Ivan ; Nakov, Preslav. / Towards automated customer support. Artificial Intelligence: Methodology, Systems, and Applications - 18th International Conference, AIMSA 2018, Proceedings. editor / Josef van Genabith ; Gennady Agre ; Thierry Declerck. Springer Verlag, 2018. pp. 48-59 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).
@inproceedings{ce3f21d3e13c4453ad6d2c7f6ffbc535,
title = "Towards automated customer support",
abstract = "Recent years have seen growing interest in conversational agents, such as chatbots, which are a very good fit for automated customer support because the domain in which they need to operate is narrow. This interest was in part inspired by recent advances in neural machine translation, esp. the rise of sequence-to-sequence (seq2seq) and attention-based models such as the Transformer, which have been applied to various other tasks and have opened new research directions in question answering, chatbots, and conversational systems. Still, in many cases, it might be feasible and even preferable to use simple information retrieval techniques. Thus, here we compare three different models: (i) a retrieval model, (ii) a sequence-to-sequence model with attention, and (iii) Transformer. Our experiments with the Twitter Customer Support Dataset, which contains over two million posts from customer support services of twenty major brands, show that the seq2seq model outperforms the other two in terms of semantics and word overlap.",
keywords = "Chatbots, Conversational agents, Customer support, IR, Seq2seq, Transformer",
author = "Momchil Hardalov and Ivan Koychev and Preslav Nakov",
year = "2018",
month = "1",
day = "1",
doi = "10.1007/978-3-319-99344-7_5",
language = "English",
isbn = "9783319993430",
series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",
publisher = "Springer Verlag",
pages = "48--59",
editor = "{van Genabith}, Josef and Gennady Agre and Thierry Declerck",
booktitle = "Artificial Intelligence",

}

TY - GEN

T1 - Towards automated customer support

AU - Hardalov, Momchil

AU - Koychev, Ivan

AU - Nakov, Preslav

PY - 2018/1/1

Y1 - 2018/1/1

N2 - Recent years have seen growing interest in conversational agents, such as chatbots, which are a very good fit for automated customer support because the domain in which they need to operate is narrow. This interest was in part inspired by recent advances in neural machine translation, esp. the rise of sequence-to-sequence (seq2seq) and attention-based models such as the Transformer, which have been applied to various other tasks and have opened new research directions in question answering, chatbots, and conversational systems. Still, in many cases, it might be feasible and even preferable to use simple information retrieval techniques. Thus, here we compare three different models: (i) a retrieval model, (ii) a sequence-to-sequence model with attention, and (iii) Transformer. Our experiments with the Twitter Customer Support Dataset, which contains over two million posts from customer support services of twenty major brands, show that the seq2seq model outperforms the other two in terms of semantics and word overlap.

AB - Recent years have seen growing interest in conversational agents, such as chatbots, which are a very good fit for automated customer support because the domain in which they need to operate is narrow. This interest was in part inspired by recent advances in neural machine translation, esp. the rise of sequence-to-sequence (seq2seq) and attention-based models such as the Transformer, which have been applied to various other tasks and have opened new research directions in question answering, chatbots, and conversational systems. Still, in many cases, it might be feasible and even preferable to use simple information retrieval techniques. Thus, here we compare three different models: (i) a retrieval model, (ii) a sequence-to-sequence model with attention, and (iii) Transformer. Our experiments with the Twitter Customer Support Dataset, which contains over two million posts from customer support services of twenty major brands, show that the seq2seq model outperforms the other two in terms of semantics and word overlap.

KW - Chatbots

KW - Conversational agents

KW - Customer support

KW - IR

KW - Seq2seq

KW - Transformer

UR - http://www.scopus.com/inward/record.url?scp=85053145700&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85053145700&partnerID=8YFLogxK

U2 - 10.1007/978-3-319-99344-7_5

DO - 10.1007/978-3-319-99344-7_5

M3 - Conference contribution

SN - 9783319993430

T3 - Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)

SP - 48

EP - 59

BT - Artificial Intelligence

A2 - van Genabith, Josef

A2 - Agre, Gennady

A2 - Declerck, Thierry

PB - Springer Verlag

ER -