Weighted matrix factorization for spoken document retrieval

Kuan Yu Chen, Hsin Min Wang, Berlin Chen, Hsin Hsi Chen

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

5 Citations (Scopus)

Abstract

Since more and more multimedia data associated with spoken documents have been made available to the public, spoken document retrieval (SDR) has become an important research subject in the past two decades. Recently, topic models have been successfully used in SDR as well as general information retrieval (IR). These models fall into two categories: probabilistic topic models (PTM) and non-probabilistic topic models (NPTM). One major difference between PTM and NPTM is that the former only takes the words occurring in a document into account, whereas the latter, such as latent semantic analysis (LSA), explicitly models all the words in the vocabulary (including both occurring and non-occurring words). We believe that the non-occurring words can provide additional information that is also useful for SDR. However, to our best knowledge, there is a dearth of work investigating the effectiveness of the non-occurring words for SDR and IR. In order to make effective use of those non-occurring words of documents for semantic analysis, we propose a weighted matrix factorization (WMF) framework, in which the impact of the non-occurring words on the semantic analysis can be modulated properly. The results of SDR experiments conducted on the TDT-2 (Topic Detection and Tracking) collection highlight the performance merits of our proposed framework when compared to several existing topic models.

Original languageEnglish
Title of host publication2013 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2013 - Proceedings
Pages8530-8534
Number of pages5
DOIs
Publication statusPublished - 2013 Oct 18
Event2013 38th IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2013 - Vancouver, BC, Canada
Duration: 2013 May 262013 May 31

Publication series

NameICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings
ISSN (Print)1520-6149

Other

Other2013 38th IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2013
CountryCanada
CityVancouver, BC
Period13/5/2613/5/31

Fingerprint

Factorization
Semantics
Information retrieval
Experiments

Keywords

  • Spoken document retrieval
  • non-occurring words
  • non-probabilistic
  • topic model

ASJC Scopus subject areas

  • Software
  • Signal Processing
  • Electrical and Electronic Engineering

Cite this

Chen, K. Y., Wang, H. M., Chen, B., & Chen, H. H. (2013). Weighted matrix factorization for spoken document retrieval. In 2013 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2013 - Proceedings (pp. 8530-8534). [6639330] (ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings). https://doi.org/10.1109/ICASSP.2013.6639330

Weighted matrix factorization for spoken document retrieval. / Chen, Kuan Yu; Wang, Hsin Min; Chen, Berlin; Chen, Hsin Hsi.

2013 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2013 - Proceedings. 2013. p. 8530-8534 6639330 (ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings).

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

Chen, KY, Wang, HM, Chen, B & Chen, HH 2013, Weighted matrix factorization for spoken document retrieval. in 2013 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2013 - Proceedings., 6639330, ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings, pp. 8530-8534, 2013 38th IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2013, Vancouver, BC, Canada, 13/5/26. https://doi.org/10.1109/ICASSP.2013.6639330
Chen KY, Wang HM, Chen B, Chen HH. Weighted matrix factorization for spoken document retrieval. In 2013 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2013 - Proceedings. 2013. p. 8530-8534. 6639330. (ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings). https://doi.org/10.1109/ICASSP.2013.6639330
Chen, Kuan Yu ; Wang, Hsin Min ; Chen, Berlin ; Chen, Hsin Hsi. / Weighted matrix factorization for spoken document retrieval. 2013 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2013 - Proceedings. 2013. pp. 8530-8534 (ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings).
@inproceedings{3947127bb4b840a0bb54d0ffdc2ef65f,
title = "Weighted matrix factorization for spoken document retrieval",
abstract = "Since more and more multimedia data associated with spoken documents have been made available to the public, spoken document retrieval (SDR) has become an important research subject in the past two decades. Recently, topic models have been successfully used in SDR as well as general information retrieval (IR). These models fall into two categories: probabilistic topic models (PTM) and non-probabilistic topic models (NPTM). One major difference between PTM and NPTM is that the former only takes the words occurring in a document into account, whereas the latter, such as latent semantic analysis (LSA), explicitly models all the words in the vocabulary (including both occurring and non-occurring words). We believe that the non-occurring words can provide additional information that is also useful for SDR. However, to our best knowledge, there is a dearth of work investigating the effectiveness of the non-occurring words for SDR and IR. In order to make effective use of those non-occurring words of documents for semantic analysis, we propose a weighted matrix factorization (WMF) framework, in which the impact of the non-occurring words on the semantic analysis can be modulated properly. The results of SDR experiments conducted on the TDT-2 (Topic Detection and Tracking) collection highlight the performance merits of our proposed framework when compared to several existing topic models.",
keywords = "Spoken document retrieval, non-occurring words, non-probabilistic, topic model",
author = "Chen, {Kuan Yu} and Wang, {Hsin Min} and Berlin Chen and Chen, {Hsin Hsi}",
year = "2013",
month = "10",
day = "18",
doi = "10.1109/ICASSP.2013.6639330",
language = "English",
isbn = "9781479903566",
series = "ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings",
pages = "8530--8534",
booktitle = "2013 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2013 - Proceedings",

}

TY - GEN

T1 - Weighted matrix factorization for spoken document retrieval

AU - Chen, Kuan Yu

AU - Wang, Hsin Min

AU - Chen, Berlin

AU - Chen, Hsin Hsi

PY - 2013/10/18

Y1 - 2013/10/18

N2 - Since more and more multimedia data associated with spoken documents have been made available to the public, spoken document retrieval (SDR) has become an important research subject in the past two decades. Recently, topic models have been successfully used in SDR as well as general information retrieval (IR). These models fall into two categories: probabilistic topic models (PTM) and non-probabilistic topic models (NPTM). One major difference between PTM and NPTM is that the former only takes the words occurring in a document into account, whereas the latter, such as latent semantic analysis (LSA), explicitly models all the words in the vocabulary (including both occurring and non-occurring words). We believe that the non-occurring words can provide additional information that is also useful for SDR. However, to our best knowledge, there is a dearth of work investigating the effectiveness of the non-occurring words for SDR and IR. In order to make effective use of those non-occurring words of documents for semantic analysis, we propose a weighted matrix factorization (WMF) framework, in which the impact of the non-occurring words on the semantic analysis can be modulated properly. The results of SDR experiments conducted on the TDT-2 (Topic Detection and Tracking) collection highlight the performance merits of our proposed framework when compared to several existing topic models.

AB - Since more and more multimedia data associated with spoken documents have been made available to the public, spoken document retrieval (SDR) has become an important research subject in the past two decades. Recently, topic models have been successfully used in SDR as well as general information retrieval (IR). These models fall into two categories: probabilistic topic models (PTM) and non-probabilistic topic models (NPTM). One major difference between PTM and NPTM is that the former only takes the words occurring in a document into account, whereas the latter, such as latent semantic analysis (LSA), explicitly models all the words in the vocabulary (including both occurring and non-occurring words). We believe that the non-occurring words can provide additional information that is also useful for SDR. However, to our best knowledge, there is a dearth of work investigating the effectiveness of the non-occurring words for SDR and IR. In order to make effective use of those non-occurring words of documents for semantic analysis, we propose a weighted matrix factorization (WMF) framework, in which the impact of the non-occurring words on the semantic analysis can be modulated properly. The results of SDR experiments conducted on the TDT-2 (Topic Detection and Tracking) collection highlight the performance merits of our proposed framework when compared to several existing topic models.

KW - Spoken document retrieval

KW - non-occurring words

KW - non-probabilistic

KW - topic model

UR - http://www.scopus.com/inward/record.url?scp=84890530840&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=84890530840&partnerID=8YFLogxK

U2 - 10.1109/ICASSP.2013.6639330

DO - 10.1109/ICASSP.2013.6639330

M3 - Conference contribution

AN - SCOPUS:84890530840

SN - 9781479903566

T3 - ICASSP, IEEE International Conference on Acoustics, Speech and Signal Processing - Proceedings

SP - 8530

EP - 8534

BT - 2013 IEEE International Conference on Acoustics, Speech, and Signal Processing, ICASSP 2013 - Proceedings

ER -