Hierarchical Graph Neural Network with Cross-Attention for Cross-Device User Matching

Ali Taghibakhshi; Mingyuan Ma; Ashwath Aithal; Onur Yilmaz; Haggai Maron; Matthew West

doi:10.1007/978-3-031-39831-5_28

Hierarchical Graph Neural Network with Cross-Attention for Cross-Device User Matching

Ali Taghibakhshi, Mingyuan Ma, Ashwath Aithal, Onur Yilmaz, Haggai Maron, Matthew West

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

Abstract

Cross-device user matching is a critical problem in numerous domains, including advertising, recommender systems, and cybersecurity. It involves identifying and linking different devices belonging to the same person, utilizing sequence logs. Previous data mining techniques have struggled to address the long-range dependencies and higher-order connections between the logs. Recently, researchers have modeled this problem as a graph problem and proposed a two-tier graph contextual embedding (TGCE) neural network architecture, which outperforms previous methods. In this paper, we propose a novel hierarchical graph neural network architecture (HGNN), which has a more computationally efficient second level design than TGCE. Furthermore, we introduce a cross-attention (Cross-Att) mechanism in our model, which improves performance by 5% compared to the state-of-the-art TGCE method.

Original language	English
Title of host publication	Big Data Analytics and Knowledge Discovery - 25th International Conference, DaWaK 2023, Proceedings
Editors	Robert Wrembel, Johann Gamper, Gabriele Kotsis, Ismail Khalil, A Min Tjoa
Publisher	Springer Science and Business Media Deutschland GmbH
Pages	303-315
Number of pages	13
ISBN (Print)	9783031398308
DOIs	https://doi.org/10.1007/978-3-031-39831-5_28
State	Published - 2023
Externally published	Yes
Event	Big Data Analytics and Knowledge Discovery - 25th International Conference, DaWaK 2023, Proceedings - Penang, Malaysia Duration: 28 Aug 2023 → 30 Aug 2023

Publication series

Name	Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)
Volume	14148 LNCS

Conference

Conference	Big Data Analytics and Knowledge Discovery - 25th International Conference, DaWaK 2023, Proceedings
Country/Territory	Malaysia
City	Penang
Period	28/08/23 → 30/08/23

Keywords

Cross-attention
Graph neural network
User matching

All Science Journal Classification (ASJC) codes

Theoretical Computer Science
General Computer Science

Access to Document

10.1007/978-3-031-39831-5_28

Cite this

Taghibakhshi, A., Ma, M., Aithal, A., Yilmaz, O., Maron, H., & West, M. (2023). Hierarchical Graph Neural Network with Cross-Attention for Cross-Device User Matching. In R. Wrembel, J. Gamper, G. Kotsis, I. Khalil, & A. M. Tjoa (Eds.), Big Data Analytics and Knowledge Discovery - 25th International Conference, DaWaK 2023, Proceedings (pp. 303-315). (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 14148 LNCS). Springer Science and Business Media Deutschland GmbH. https://doi.org/10.1007/978-3-031-39831-5_28

Taghibakhshi, Ali ; Ma, Mingyuan ; Aithal, Ashwath et al. / Hierarchical Graph Neural Network with Cross-Attention for Cross-Device User Matching. Big Data Analytics and Knowledge Discovery - 25th International Conference, DaWaK 2023, Proceedings. editor / Robert Wrembel ; Johann Gamper ; Gabriele Kotsis ; Ismail Khalil ; A Min Tjoa. Springer Science and Business Media Deutschland GmbH, 2023. pp. 303-315 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)).

@inproceedings{3520755e596443649c5c8aa181e625ac,

title = "Hierarchical Graph Neural Network with Cross-Attention for Cross-Device User Matching",

abstract = "Cross-device user matching is a critical problem in numerous domains, including advertising, recommender systems, and cybersecurity. It involves identifying and linking different devices belonging to the same person, utilizing sequence logs. Previous data mining techniques have struggled to address the long-range dependencies and higher-order connections between the logs. Recently, researchers have modeled this problem as a graph problem and proposed a two-tier graph contextual embedding (TGCE) neural network architecture, which outperforms previous methods. In this paper, we propose a novel hierarchical graph neural network architecture (HGNN), which has a more computationally efficient second level design than TGCE. Furthermore, we introduce a cross-attention (Cross-Att) mechanism in our model, which improves performance by 5\% compared to the state-of-the-art TGCE method.",

keywords = "Cross-attention, Graph neural network, User matching",

author = "Ali Taghibakhshi and Mingyuan Ma and Ashwath Aithal and Onur Yilmaz and Haggai Maron and Matthew West",

note = "Publisher Copyright: {\textcopyright} 2023, The Author(s), under exclusive license to Springer Nature Switzerland AG.; Big Data Analytics and Knowledge Discovery - 25th International Conference, DaWaK 2023, Proceedings ; Conference date: 28-08-2023 Through 30-08-2023",

year = "2023",

doi = "10.1007/978-3-031-39831-5\_28",

language = "الإنجليزيّة",

isbn = "9783031398308",

series = "Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)",

publisher = "Springer Science and Business Media Deutschland GmbH",

pages = "303--315",

editor = "Robert Wrembel and Johann Gamper and Gabriele Kotsis and Ismail Khalil and Tjoa, \{A Min\}",

booktitle = "Big Data Analytics and Knowledge Discovery - 25th International Conference, DaWaK 2023, Proceedings",

address = "ألمانيا",

}

Taghibakhshi, A, Ma, M, Aithal, A, Yilmaz, O, Maron, H & West, M 2023, Hierarchical Graph Neural Network with Cross-Attention for Cross-Device User Matching. in R Wrembel, J Gamper, G Kotsis, I Khalil & AM Tjoa (eds), Big Data Analytics and Knowledge Discovery - 25th International Conference, DaWaK 2023, Proceedings. Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics), vol. 14148 LNCS, Springer Science and Business Media Deutschland GmbH, pp. 303-315, Big Data Analytics and Knowledge Discovery - 25th International Conference, DaWaK 2023, Proceedings, Penang, Malaysia, 28/08/23. https://doi.org/10.1007/978-3-031-39831-5_28

Hierarchical Graph Neural Network with Cross-Attention for Cross-Device User Matching. / Taghibakhshi, Ali; Ma, Mingyuan; Aithal, Ashwath et al.
Big Data Analytics and Knowledge Discovery - 25th International Conference, DaWaK 2023, Proceedings. ed. / Robert Wrembel; Johann Gamper; Gabriele Kotsis; Ismail Khalil; A Min Tjoa. Springer Science and Business Media Deutschland GmbH, 2023. p. 303-315 (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics); Vol. 14148 LNCS).

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

TY - GEN

T1 - Hierarchical Graph Neural Network with Cross-Attention for Cross-Device User Matching

AU - Taghibakhshi, Ali

AU - Ma, Mingyuan

AU - Aithal, Ashwath

AU - Yilmaz, Onur

AU - Maron, Haggai

AU - West, Matthew

PY - 2023

Y1 - 2023

N2 - Cross-device user matching is a critical problem in numerous domains, including advertising, recommender systems, and cybersecurity. It involves identifying and linking different devices belonging to the same person, utilizing sequence logs. Previous data mining techniques have struggled to address the long-range dependencies and higher-order connections between the logs. Recently, researchers have modeled this problem as a graph problem and proposed a two-tier graph contextual embedding (TGCE) neural network architecture, which outperforms previous methods. In this paper, we propose a novel hierarchical graph neural network architecture (HGNN), which has a more computationally efficient second level design than TGCE. Furthermore, we introduce a cross-attention (Cross-Att) mechanism in our model, which improves performance by 5% compared to the state-of-the-art TGCE method.

AB - Cross-device user matching is a critical problem in numerous domains, including advertising, recommender systems, and cybersecurity. It involves identifying and linking different devices belonging to the same person, utilizing sequence logs. Previous data mining techniques have struggled to address the long-range dependencies and higher-order connections between the logs. Recently, researchers have modeled this problem as a graph problem and proposed a two-tier graph contextual embedding (TGCE) neural network architecture, which outperforms previous methods. In this paper, we propose a novel hierarchical graph neural network architecture (HGNN), which has a more computationally efficient second level design than TGCE. Furthermore, we introduce a cross-attention (Cross-Att) mechanism in our model, which improves performance by 5% compared to the state-of-the-art TGCE method.

KW - Cross-attention

KW - Graph neural network

KW - User matching

UR - http://www.scopus.com/inward/record.url?scp=85172409302&partnerID=8YFLogxK

U2 - 10.1007/978-3-031-39831-5_28

DO - 10.1007/978-3-031-39831-5_28

M3 - منشور من مؤتمر

SN - 9783031398308

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

SP - 303

EP - 315

BT - Big Data Analytics and Knowledge Discovery - 25th International Conference, DaWaK 2023, Proceedings

A2 - Wrembel, Robert

A2 - Gamper, Johann

A2 - Kotsis, Gabriele

A2 - Khalil, Ismail

A2 - Tjoa, A Min

PB - Springer Science and Business Media Deutschland GmbH

T2 - Big Data Analytics and Knowledge Discovery - 25th International Conference, DaWaK 2023, Proceedings

Y2 - 28 August 2023 through 30 August 2023

ER -

Taghibakhshi A, Ma M, Aithal A, Yilmaz O, Maron H, West M. Hierarchical Graph Neural Network with Cross-Attention for Cross-Device User Matching. In Wrembel R, Gamper J, Kotsis G, Khalil I, Tjoa AM, editors, Big Data Analytics and Knowledge Discovery - 25th International Conference, DaWaK 2023, Proceedings. Springer Science and Business Media Deutschland GmbH. 2023. p. 303-315. (Lecture Notes in Computer Science (including subseries Lecture Notes in Artificial Intelligence and Lecture Notes in Bioinformatics)). doi: 10.1007/978-3-031-39831-5_28

Hierarchical Graph Neural Network with Cross-Attention for Cross-Device User Matching

Abstract

Publication series

Conference

Keywords

All Science Journal Classification (ASJC) codes

Access to Document

Other files and links

Fingerprint

Cite this