Copula Guided Parallel Gibbs Sampling for Nonparametric and Coherent Topic Discovery (Extended Abstract)

Lihui LIN; Yanghui RAO; Haoran XIE; Raymond Y. K. LAU; Jian YIN; Fu Lee WANG; Qing LI

doi:10.1109/ICDE55515.2023.00338

Copula Guided Parallel Gibbs Sampling for Nonparametric and Coherent Topic Discovery (Extended Abstract)

Lihui LIN, Yanghui RAO, Haoran XIE, Raymond Y. K. LAU, Jian YIN, Fu Lee WANG, Qing LI

Department of Computing and Decision Sciences

Research output: Book Chapters | Papers in Conference Proceedings › Conference (Extended Abstracts) › peer-review

Abstract

In terms of the generative process, the Gamma-Gamma-Poisson Process (G2PP) is equivalent to the nonparametric topic model of Hierarchical Dirichlet Process (HDP). Considering the high computational cost of estimating parameters in HDP, a parallel G2PP was developed to generate topics efficiently via multi-threading. Unfortunately, the above model needs to predefine the number of topics. To address this issue, we first propose a Topic Self-Adaptive Model (TSAM) for nonparametric and parallel topic discovery. In TSAM, a monitor-executor mechanism is developed to manage the global topic information using a hierarchical structure of threads. Based on the apparatus of copulas, we further extend our TSAM to TSAMcop for coherent topic modeling by exploiting a copula guided parallel Gibbs sampling algorithm. Extensive experiments validate the effectiveness of both TSAM and TSAMcop.

Original language	English
Title of host publication	Proceedings - 2023 IEEE 39th International Conference on Data Engineering, ICDE 2023
Pages	3823-3824
Number of pages	2
ISBN (Electronic)	9798350322279
DOIs	https://doi.org/10.1109/ICDE55515.2023.00338
Publication status	Published - 26 Jul 2023
Event	2023 IEEE 39th International Conference on Data Engineering (ICDE) - Anaheim, United States Duration: 3 Apr 2023 → 7 Apr 2023 https://icde2023.ics.uci.edu/

Publication series

Name	Proceedings - International Conference on Data Engineering
Volume	2023-April
ISSN (Print)	1084-4627

Conference

Conference	2023 IEEE 39th International Conference on Data Engineering (ICDE)
Country/Territory	United States
City	Anaheim
Period	3/04/23 → 7/04/23
Internet address	https://icde2023.ics.uci.edu/

Bibliographical note

Publisher Copyright:
© 2023 IEEE.

Funding

This work was supported by a grant from the Research Grants Council of the Hong Kong Special Administrative Region, China (UGC/FDS16/E01/19), the National Natural Science Foundation of China (61972426), the Direct Grant (DR23B2) and the Faculty Research Grant (DB23A3) of Lingnan University, Hong Kong, a grant from the Research Grants Council of the HKSAR, China (Project: CityU 11507219), and a grant from the City University of Hong Kong SRG (Project: 7005780).

Keywords

copulas
parallel gibbs sampling
topic modelling

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

Access to Document

10.1109/ICDE55515.2023.00338

https://ieeexplore.ieee.org/document/10184894/

Cite this

LIN, L., RAO, Y., XIE, H., LAU, R. Y. K., YIN, J., WANG, F. L., & LI, Q. (2023). Copula Guided Parallel Gibbs Sampling for Nonparametric and Coherent Topic Discovery (Extended Abstract). In Proceedings - 2023 IEEE 39th International Conference on Data Engineering, ICDE 2023 (pp. 3823-3824). (Proceedings - International Conference on Data Engineering; Vol. 2023-April). https://doi.org/10.1109/ICDE55515.2023.00338

@inbook{c6e75bd64f4140e7b69af37387aa2393,

title = "Copula Guided Parallel Gibbs Sampling for Nonparametric and Coherent Topic Discovery (Extended Abstract)",

abstract = "In terms of the generative process, the Gamma-Gamma-Poisson Process (G2PP) is equivalent to the nonparametric topic model of Hierarchical Dirichlet Process (HDP). Considering the high computational cost of estimating parameters in HDP, a parallel G2PP was developed to generate topics efficiently via multi-threading. Unfortunately, the above model needs to predefine the number of topics. To address this issue, we first propose a Topic Self-Adaptive Model (TSAM) for nonparametric and parallel topic discovery. In TSAM, a monitor-executor mechanism is developed to manage the global topic information using a hierarchical structure of threads. Based on the apparatus of copulas, we further extend our TSAM to TSAMcop for coherent topic modeling by exploiting a copula guided parallel Gibbs sampling algorithm. Extensive experiments validate the effectiveness of both TSAM and TSAMcop.",

keywords = "copulas, parallel gibbs sampling, topic modelling",

author = "Lihui LIN and Yanghui RAO and Haoran XIE and LAU, {Raymond Y. K.} and Jian YIN and WANG, {Fu Lee} and Qing LI",

note = "Publisher Copyright: {\textcopyright} 2023 IEEE.; 2023 IEEE 39th International Conference on Data Engineering (ICDE) ; Conference date: 03-04-2023 Through 07-04-2023",

year = "2023",

month = jul,

day = "26",

doi = "10.1109/ICDE55515.2023.00338",

language = "English",

isbn = "9798350322279",

series = "Proceedings - International Conference on Data Engineering",

pages = "3823--3824",

booktitle = "Proceedings - 2023 IEEE 39th International Conference on Data Engineering, ICDE 2023",

url = "https://icde2023.ics.uci.edu/",

}

LIN, L, RAO, Y, XIE, H, LAU, RYK, YIN, J, WANG, FL & LI, Q 2023, Copula Guided Parallel Gibbs Sampling for Nonparametric and Coherent Topic Discovery (Extended Abstract). in Proceedings - 2023 IEEE 39th International Conference on Data Engineering, ICDE 2023. Proceedings - International Conference on Data Engineering, vol. 2023-April, pp. 3823-3824, 2023 IEEE 39th International Conference on Data Engineering (ICDE), Anaheim, California, United States, 3/04/23. https://doi.org/10.1109/ICDE55515.2023.00338

Copula Guided Parallel Gibbs Sampling for Nonparametric and Coherent Topic Discovery (Extended Abstract). / LIN, Lihui; RAO, Yanghui; XIE, Haoran et al.
Proceedings - 2023 IEEE 39th International Conference on Data Engineering, ICDE 2023. 2023. p. 3823-3824 (Proceedings - International Conference on Data Engineering; Vol. 2023-April).

Research output: Book Chapters | Papers in Conference Proceedings › Conference (Extended Abstracts) › peer-review

TY - CHAP

T1 - Copula Guided Parallel Gibbs Sampling for Nonparametric and Coherent Topic Discovery (Extended Abstract)

AU - LIN, Lihui

AU - RAO, Yanghui

AU - XIE, Haoran

AU - LAU, Raymond Y. K.

AU - YIN, Jian

AU - WANG, Fu Lee

AU - LI, Qing

PY - 2023/7/26

Y1 - 2023/7/26

N2 - In terms of the generative process, the Gamma-Gamma-Poisson Process (G2PP) is equivalent to the nonparametric topic model of Hierarchical Dirichlet Process (HDP). Considering the high computational cost of estimating parameters in HDP, a parallel G2PP was developed to generate topics efficiently via multi-threading. Unfortunately, the above model needs to predefine the number of topics. To address this issue, we first propose a Topic Self-Adaptive Model (TSAM) for nonparametric and parallel topic discovery. In TSAM, a monitor-executor mechanism is developed to manage the global topic information using a hierarchical structure of threads. Based on the apparatus of copulas, we further extend our TSAM to TSAMcop for coherent topic modeling by exploiting a copula guided parallel Gibbs sampling algorithm. Extensive experiments validate the effectiveness of both TSAM and TSAMcop.

AB - In terms of the generative process, the Gamma-Gamma-Poisson Process (G2PP) is equivalent to the nonparametric topic model of Hierarchical Dirichlet Process (HDP). Considering the high computational cost of estimating parameters in HDP, a parallel G2PP was developed to generate topics efficiently via multi-threading. Unfortunately, the above model needs to predefine the number of topics. To address this issue, we first propose a Topic Self-Adaptive Model (TSAM) for nonparametric and parallel topic discovery. In TSAM, a monitor-executor mechanism is developed to manage the global topic information using a hierarchical structure of threads. Based on the apparatus of copulas, we further extend our TSAM to TSAMcop for coherent topic modeling by exploiting a copula guided parallel Gibbs sampling algorithm. Extensive experiments validate the effectiveness of both TSAM and TSAMcop.

KW - copulas

KW - parallel gibbs sampling

KW - topic modelling

UR - http://www.scopus.com/inward/record.url?scp=85167677165&partnerID=8YFLogxK

U2 - 10.1109/ICDE55515.2023.00338

DO - 10.1109/ICDE55515.2023.00338

M3 - Conference (Extended Abstracts)

SN - 9798350322279

T3 - Proceedings - International Conference on Data Engineering

SP - 3823

EP - 3824

BT - Proceedings - 2023 IEEE 39th International Conference on Data Engineering, ICDE 2023

T2 - 2023 IEEE 39th International Conference on Data Engineering (ICDE)

Y2 - 3 April 2023 through 7 April 2023

ER -

Copula Guided Parallel Gibbs Sampling for Nonparametric and Coherent Topic Discovery (Extended Abstract)

Abstract

Publication series

Conference

Bibliographical note

Funding

Keywords

UN SDGs

Access to Document

Other files and links

Fingerprint

Cite this