The Context-Dependent Additive Recurrent Neural Net

Hung Quan Tran, Tuan Manh Lai, Gholamreza Haffari, Ingrid Zukerman, Trung Bui, Hung Bui

Research output: Chapter in Book/Report/Conference proceedingConference PaperResearchpeer-review

Abstract

Contextual sequence mapping is one of the fundamental problems in Natural Language Processing. Instead of relying solely on the information presented in a text, the learning agents have access to a strong external signal given to assist the learning process. In this paper, we propose a novel family of Recurrent Neural Network unit: the Context-dependent Additive Recurrent Neural Network (CARNN) that is designed specifically to leverage this external signal. The experimental results on pubic datasets in the dialog problem (Babi dialog Task 6 and Frame), contextual language model (Switchboard and Penn Discourse Tree Bank) and question answering (TrecQA) show that our novel CARNN-based architectures outperform previous methods.
Original languageEnglish
Title of host publicationConference of the North American Chapter of the Association for Computational Linguistics
Subtitle of host publicationHuman Language Technologies (NAACL HLT 2018)
EditorsHeng Ji, Amanda Stent
Place of PublicationRed Hook NY USA
PublisherAssociation for Information Systems
Pages1274-1283
Number of pages10
Volume1
ISBN (Electronic)9781510863460
Publication statusPublished - 2018
EventNorth American Association for Computational Linguistics 2018: Human Language Technologies - Hyatt Regency, New Orleans, United States of America
Duration: 1 Jun 20186 Jun 2018
Conference number: 16th
http://naacl2018.org/

Conference

ConferenceNorth American Association for Computational Linguistics 2018
Abbreviated titleNAACL HLT 2018
CountryUnited States of America
CityNew Orleans
Period1/06/186/06/18
Internet address

Keywords

  • sequence mapping

Cite this

Tran, H. Q., Lai, T. M., Haffari, G., Zukerman, I., Bui, T., & Bui, H. (2018). The Context-Dependent Additive Recurrent Neural Net. In H. Ji, & A. Stent (Eds.), Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL HLT 2018) (Vol. 1, pp. 1274-1283). Red Hook NY USA: Association for Information Systems.
Tran, Hung Quan ; Lai, Tuan Manh ; Haffari, Gholamreza ; Zukerman, Ingrid ; Bui, Trung ; Bui, Hung. / The Context-Dependent Additive Recurrent Neural Net. Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL HLT 2018). editor / Heng Ji ; Amanda Stent. Vol. 1 Red Hook NY USA : Association for Information Systems, 2018. pp. 1274-1283
@inproceedings{f37759ff20a84d70be666da4b8cb8520,
title = "The Context-Dependent Additive Recurrent Neural Net",
abstract = "Contextual sequence mapping is one of the fundamental problems in Natural Language Processing. Instead of relying solely on the information presented in a text, the learning agents have access to a strong external signal given to assist the learning process. In this paper, we propose a novel family of Recurrent Neural Network unit: the Context-dependent Additive Recurrent Neural Network (CARNN) that is designed specifically to leverage this external signal. The experimental results on pubic datasets in the dialog problem (Babi dialog Task 6 and Frame), contextual language model (Switchboard and Penn Discourse Tree Bank) and question answering (TrecQA) show that our novel CARNN-based architectures outperform previous methods.",
keywords = "sequence mapping",
author = "Tran, {Hung Quan} and Lai, {Tuan Manh} and Gholamreza Haffari and Ingrid Zukerman and Trung Bui and Hung Bui",
year = "2018",
language = "English",
volume = "1",
pages = "1274--1283",
editor = "Heng Ji and Amanda Stent",
booktitle = "Conference of the North American Chapter of the Association for Computational Linguistics",
publisher = "Association for Information Systems",
address = "United States of America",

}

Tran, HQ, Lai, TM, Haffari, G, Zukerman, I, Bui, T & Bui, H 2018, The Context-Dependent Additive Recurrent Neural Net. in H Ji & A Stent (eds), Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL HLT 2018). vol. 1, Association for Information Systems, Red Hook NY USA, pp. 1274-1283, North American Association for Computational Linguistics 2018, New Orleans, United States of America, 1/06/18.

The Context-Dependent Additive Recurrent Neural Net. / Tran, Hung Quan; Lai, Tuan Manh; Haffari, Gholamreza; Zukerman, Ingrid; Bui, Trung; Bui, Hung.

Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL HLT 2018). ed. / Heng Ji; Amanda Stent. Vol. 1 Red Hook NY USA : Association for Information Systems, 2018. p. 1274-1283.

Research output: Chapter in Book/Report/Conference proceedingConference PaperResearchpeer-review

TY - GEN

T1 - The Context-Dependent Additive Recurrent Neural Net

AU - Tran, Hung Quan

AU - Lai, Tuan Manh

AU - Haffari, Gholamreza

AU - Zukerman, Ingrid

AU - Bui, Trung

AU - Bui, Hung

PY - 2018

Y1 - 2018

N2 - Contextual sequence mapping is one of the fundamental problems in Natural Language Processing. Instead of relying solely on the information presented in a text, the learning agents have access to a strong external signal given to assist the learning process. In this paper, we propose a novel family of Recurrent Neural Network unit: the Context-dependent Additive Recurrent Neural Network (CARNN) that is designed specifically to leverage this external signal. The experimental results on pubic datasets in the dialog problem (Babi dialog Task 6 and Frame), contextual language model (Switchboard and Penn Discourse Tree Bank) and question answering (TrecQA) show that our novel CARNN-based architectures outperform previous methods.

AB - Contextual sequence mapping is one of the fundamental problems in Natural Language Processing. Instead of relying solely on the information presented in a text, the learning agents have access to a strong external signal given to assist the learning process. In this paper, we propose a novel family of Recurrent Neural Network unit: the Context-dependent Additive Recurrent Neural Network (CARNN) that is designed specifically to leverage this external signal. The experimental results on pubic datasets in the dialog problem (Babi dialog Task 6 and Frame), contextual language model (Switchboard and Penn Discourse Tree Bank) and question answering (TrecQA) show that our novel CARNN-based architectures outperform previous methods.

KW - sequence mapping

M3 - Conference Paper

VL - 1

SP - 1274

EP - 1283

BT - Conference of the North American Chapter of the Association for Computational Linguistics

A2 - Ji, Heng

A2 - Stent, Amanda

PB - Association for Information Systems

CY - Red Hook NY USA

ER -

Tran HQ, Lai TM, Haffari G, Zukerman I, Bui T, Bui H. The Context-Dependent Additive Recurrent Neural Net. In Ji H, Stent A, editors, Conference of the North American Chapter of the Association for Computational Linguistics: Human Language Technologies (NAACL HLT 2018). Vol. 1. Red Hook NY USA: Association for Information Systems. 2018. p. 1274-1283