Enhanced Bert-Based Ranking Models for Spoken Document Retrieval

Hsiao Yun Lin; Tien Hong Lo; Berlin Chen

doi:10.1109/ASRU46091.2019.9003890

Enhanced Bert-Based Ranking Models for Spoken Document Retrieval

Hsiao Yun Lin, Tien Hong Lo, Berlin Chen

Department of Computer Science and Information Engineering

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

8 Citations (Scopus)

Abstract

The Bidirectional Encoder Representations from Transformers (BERT) model has recently achieved record-breaking success on many natural language processing (NLP) tasks such as question answering and language understanding. However, relatively little work has been done on ad-hoc information retrieval (IR), especially for spoken document retrieval (SDR). This paper adopts and extends BERT for SDR, while its contributions are at least three-fold. First, we augment BERT with extra language features such as unigram and inverse document frequency (IDF) statistics to make it more applicable to SDR. Second, we also explore the incorporation of confidence scores into document representations to see if they could help alleviate the negative effects resulting from imperfect automatic speech recognition (ASR). Third, we conduct a comprehensive set of experiments to compare our BERT-based ranking methods with other state-of-The-Art ones and investigate the synergy effect of them as well.

Original language	English
Title of host publication	2019 IEEE Automatic Speech Recognition and Understanding Workshop, ASRU 2019 - Proceedings
Publisher	Institute of Electrical and Electronics Engineers Inc.
Pages	601-606
Number of pages	6
ISBN (Electronic)	9781728103068
DOIs	https://doi.org/10.1109/ASRU46091.2019.9003890
Publication status	Published - 2019 Dec
Event	2019 IEEE Automatic Speech Recognition and Understanding Workshop, ASRU 2019 - Singapore, Singapore Duration: 2019 Dec 15 → 2019 Dec 18

Publication series

Name	2019 IEEE Automatic Speech Recognition and Understanding Workshop, ASRU 2019 - Proceedings

Conference

Conference	2019 IEEE Automatic Speech Recognition and Understanding Workshop, ASRU 2019
Country/Territory	Singapore
City	Singapore
Period	2019/12/15 → 2019/12/18

Keywords

BERT
Spoken document retrieval
information retrieval
model augmentation
speech recognition

ASJC Scopus subject areas

Computer Networks and Communications
Signal Processing
Linguistics and Language
Communication

Access to Document

10.1109/ASRU46091.2019.9003890

Cite this

Lin, H. Y., Lo, T. H., & Chen, B. (2019). Enhanced Bert-Based Ranking Models for Spoken Document Retrieval. In 2019 IEEE Automatic Speech Recognition and Understanding Workshop, ASRU 2019 - Proceedings (pp. 601-606). Article 9003890 (2019 IEEE Automatic Speech Recognition and Understanding Workshop, ASRU 2019 - Proceedings). Institute of Electrical and Electronics Engineers Inc.. https://doi.org/10.1109/ASRU46091.2019.9003890

Enhanced Bert-Based Ranking Models for Spoken Document Retrieval. / Lin, Hsiao Yun; Lo, Tien Hong; Chen, Berlin.
2019 IEEE Automatic Speech Recognition and Understanding Workshop, ASRU 2019 - Proceedings. Institute of Electrical and Electronics Engineers Inc., 2019. p. 601-606 9003890 (2019 IEEE Automatic Speech Recognition and Understanding Workshop, ASRU 2019 - Proceedings).

Research output: Chapter in Book/Report/Conference proceeding › Conference contribution

Lin, HY, Lo, TH & Chen, B 2019, Enhanced Bert-Based Ranking Models for Spoken Document Retrieval. in 2019 IEEE Automatic Speech Recognition and Understanding Workshop, ASRU 2019 - Proceedings., 9003890, 2019 IEEE Automatic Speech Recognition and Understanding Workshop, ASRU 2019 - Proceedings, Institute of Electrical and Electronics Engineers Inc., pp. 601-606, 2019 IEEE Automatic Speech Recognition and Understanding Workshop, ASRU 2019, Singapore, Singapore, 2019/12/15. https://doi.org/10.1109/ASRU46091.2019.9003890

Lin HY, Lo TH, Chen B. Enhanced Bert-Based Ranking Models for Spoken Document Retrieval. In 2019 IEEE Automatic Speech Recognition and Understanding Workshop, ASRU 2019 - Proceedings. Institute of Electrical and Electronics Engineers Inc. 2019. p. 601-606. 9003890. (2019 IEEE Automatic Speech Recognition and Understanding Workshop, ASRU 2019 - Proceedings). doi: 10.1109/ASRU46091.2019.9003890

Lin, Hsiao Yun ; Lo, Tien Hong ; Chen, Berlin. / Enhanced Bert-Based Ranking Models for Spoken Document Retrieval. 2019 IEEE Automatic Speech Recognition and Understanding Workshop, ASRU 2019 - Proceedings. Institute of Electrical and Electronics Engineers Inc., 2019. pp. 601-606 (2019 IEEE Automatic Speech Recognition and Understanding Workshop, ASRU 2019 - Proceedings).

@inproceedings{db59fdf51c044c39bbdd66d0c6acecf0,

title = "Enhanced Bert-Based Ranking Models for Spoken Document Retrieval",

abstract = "The Bidirectional Encoder Representations from Transformers (BERT) model has recently achieved record-breaking success on many natural language processing (NLP) tasks such as question answering and language understanding. However, relatively little work has been done on ad-hoc information retrieval (IR), especially for spoken document retrieval (SDR). This paper adopts and extends BERT for SDR, while its contributions are at least three-fold. First, we augment BERT with extra language features such as unigram and inverse document frequency (IDF) statistics to make it more applicable to SDR. Second, we also explore the incorporation of confidence scores into document representations to see if they could help alleviate the negative effects resulting from imperfect automatic speech recognition (ASR). Third, we conduct a comprehensive set of experiments to compare our BERT-based ranking methods with other state-of-The-Art ones and investigate the synergy effect of them as well.",

keywords = "BERT, Spoken document retrieval, information retrieval, model augmentation, speech recognition",

author = "Lin, {Hsiao Yun} and Lo, {Tien Hong} and Berlin Chen",

note = "Publisher Copyright: {\textcopyright} 2019 IEEE.; 2019 IEEE Automatic Speech Recognition and Understanding Workshop, ASRU 2019 ; Conference date: 15-12-2019 Through 18-12-2019",

year = "2019",

month = dec,

doi = "10.1109/ASRU46091.2019.9003890",

language = "English",

series = "2019 IEEE Automatic Speech Recognition and Understanding Workshop, ASRU 2019 - Proceedings",

publisher = "Institute of Electrical and Electronics Engineers Inc.",

pages = "601--606",

booktitle = "2019 IEEE Automatic Speech Recognition and Understanding Workshop, ASRU 2019 - Proceedings",

}

TY - GEN

T1 - Enhanced Bert-Based Ranking Models for Spoken Document Retrieval

AU - Lin, Hsiao Yun

AU - Lo, Tien Hong

AU - Chen, Berlin

PY - 2019/12

Y1 - 2019/12

N2 - The Bidirectional Encoder Representations from Transformers (BERT) model has recently achieved record-breaking success on many natural language processing (NLP) tasks such as question answering and language understanding. However, relatively little work has been done on ad-hoc information retrieval (IR), especially for spoken document retrieval (SDR). This paper adopts and extends BERT for SDR, while its contributions are at least three-fold. First, we augment BERT with extra language features such as unigram and inverse document frequency (IDF) statistics to make it more applicable to SDR. Second, we also explore the incorporation of confidence scores into document representations to see if they could help alleviate the negative effects resulting from imperfect automatic speech recognition (ASR). Third, we conduct a comprehensive set of experiments to compare our BERT-based ranking methods with other state-of-The-Art ones and investigate the synergy effect of them as well.

AB - The Bidirectional Encoder Representations from Transformers (BERT) model has recently achieved record-breaking success on many natural language processing (NLP) tasks such as question answering and language understanding. However, relatively little work has been done on ad-hoc information retrieval (IR), especially for spoken document retrieval (SDR). This paper adopts and extends BERT for SDR, while its contributions are at least three-fold. First, we augment BERT with extra language features such as unigram and inverse document frequency (IDF) statistics to make it more applicable to SDR. Second, we also explore the incorporation of confidence scores into document representations to see if they could help alleviate the negative effects resulting from imperfect automatic speech recognition (ASR). Third, we conduct a comprehensive set of experiments to compare our BERT-based ranking methods with other state-of-The-Art ones and investigate the synergy effect of them as well.

KW - BERT

KW - Spoken document retrieval

KW - information retrieval

KW - model augmentation

KW - speech recognition

UR - http://www.scopus.com/inward/record.url?scp=85081595716&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85081595716&partnerID=8YFLogxK

U2 - 10.1109/ASRU46091.2019.9003890

DO - 10.1109/ASRU46091.2019.9003890

M3 - Conference contribution

AN - SCOPUS:85081595716

T3 - 2019 IEEE Automatic Speech Recognition and Understanding Workshop, ASRU 2019 - Proceedings

SP - 601

EP - 606

BT - 2019 IEEE Automatic Speech Recognition and Understanding Workshop, ASRU 2019 - Proceedings

PB - Institute of Electrical and Electronics Engineers Inc.

T2 - 2019 IEEE Automatic Speech Recognition and Understanding Workshop, ASRU 2019

Y2 - 15 December 2019 through 18 December 2019

ER -

Enhanced Bert-Based Ranking Models for Spoken Document Retrieval

Abstract

Publication series

Conference

Keywords

ASJC Scopus subject areas

Access to Document

Other files and links

Fingerprint

Cite this