MEET-Sepsis: Multi-Endogenous-View Enhanced Time-Series Representation Learning for Early Sepsis Prediction

Zexi Tan; Tao Xie; Binbin Sun; Xiang Zhang; Yiqun Zhang; Yiu Ming Cheung

doi:10.1007/978-981-95-7084-3_47

MEET-Sepsis: Multi-Endogenous-View Enhanced Time-Series Representation Learning for Early Sepsis Prediction

Zexi Tan (Co-first author)
, Tao Xie (Co-first author)
, Binbin Sun
, Xiang Zhang
, Yiqun Zhang^*
, Yiu Ming Cheung^*

^*Corresponding author for this work

Department of Computer Science

Research output: Chapter in book/report/conference proceeding › Conference proceeding › peer-review

Abstract

Sepsis is a life-threatening infectious syndrome associated with high mortality in intensive care units (ICUs). Early and accurate sepsis prediction (SP) is critical for timely intervention, yet remains challenging due to subtle early manifestations and rapidly escalating mortality. While AI has improved SP efficiency, existing methods struggle to capture weak early temporal signals. This paper introduces a Multi-Endogenous-view Representation Enhancement (MERE) mechanism to construct enriched feature views, coupled with a Cascaded Dual-convolution Time-series Attention (CDTA) module for multi-scale temporal representation learning. The proposed MEET-Sepsis framework achieves competitive prediction accuracy using only 20% of the ICU monitoring time required by SOTA methods, significantly advancing early SP. Extensive validation confirms its efficacy. Code is available at: https://github.com/yueliangy/MEET-Sepsis.

Original language	English
Title of host publication	PRICAI 2025: Trends in Artificial Intelligence
Subtitle of host publication	22nd Pacific Rim International Conference on Artificial Intelligence, PRICAI 2025, Wellington, New Zealand, November 17–21, 2025, Proceedings, Part V
Editors	Yi Mei, Chao Qian, Quan Bai, Bing Xue, Sankalp Khanna
Publisher	Springer
Pages	678-686
Number of pages	9
ISBN (Electronic)	9789819570843
ISBN (Print)	9789819570836
DOIs	https://doi.org/10.1007/978-981-95-7084-3_47
Publication status	Published - 8 May 2026
Event	22nd Pacific Rim International Conference on Artificial Intelligence, PRICAI 2025 - Wellington, New Zealand Duration: 17 Nov 2025 → 21 Nov 2025 https://link.springer.com/book/10.1007/978-981-95-7084-3 (Conference proceedings)

Publication series

Name	Lecture Notes in Computer Science
Volume	16455
ISSN (Print)	0302-9743
ISSN (Electronic)	1611-3349

Name	Lecture Notes in Artificial Intelligence
ISSN (Print)	2945-9133
ISSN (Electronic)	2945-9141

Name	PRICAI: Pacific Rim International Conference on Artificial Intelligence

Conference

Conference	22nd Pacific Rim International Conference on Artificial Intelligence, PRICAI 2025
Abbreviated title	PRICAI 2025
Country/Territory	New Zealand
City	Wellington
Period	17/11/25 → 21/11/25
Internet address	https://link.springer.com/book/10.1007/978-981-95-7084-3 (Conference proceedings)

UN SDGs

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

SDG 3 Good Health and Well-being

User-Defined Keywords

Attention mechanism
Classification
Early sepsis prediction
Multi-view learning
Supervised learning
Time-series

Access to Document

10.1007/978-981-95-7084-3_47

Cite this

Tan, Z., Xie, T., Sun, B., Zhang, X., Zhang, Y., & Cheung, Y. M. (2026). MEET-Sepsis: Multi-Endogenous-View Enhanced Time-Series Representation Learning for Early Sepsis Prediction. In Y. Mei, C. Qian, Q. Bai, B. Xue, & S. Khanna (Eds.), PRICAI 2025: Trends in Artificial Intelligence: 22nd Pacific Rim International Conference on Artificial Intelligence, PRICAI 2025, Wellington, New Zealand, November 17–21, 2025, Proceedings, Part V (pp. 678-686). (Lecture Notes in Computer Science; Vol. 16455), (Lecture Notes in Artificial Intelligence), (PRICAI: Pacific Rim International Conference on Artificial Intelligence). Springer. https://doi.org/10.1007/978-981-95-7084-3_47

Tan, Zexi ; Xie, Tao ; Sun, Binbin et al. / MEET-Sepsis: Multi-Endogenous-View Enhanced Time-Series Representation Learning for Early Sepsis Prediction. PRICAI 2025: Trends in Artificial Intelligence: 22nd Pacific Rim International Conference on Artificial Intelligence, PRICAI 2025, Wellington, New Zealand, November 17–21, 2025, Proceedings, Part V. editor / Yi Mei ; Chao Qian ; Quan Bai ; Bing Xue ; Sankalp Khanna. Springer, 2026. pp. 678-686 (Lecture Notes in Computer Science). (Lecture Notes in Artificial Intelligence). (PRICAI: Pacific Rim International Conference on Artificial Intelligence).

@inproceedings{be28701a9ad14d01b72b5032aa5ada83,

title = "MEET-Sepsis: Multi-Endogenous-View Enhanced Time-Series Representation Learning for Early Sepsis Prediction",

abstract = "Sepsis is a life-threatening infectious syndrome associated with high mortality in intensive care units (ICUs). Early and accurate sepsis prediction (SP) is critical for timely intervention, yet remains challenging due to subtle early manifestations and rapidly escalating mortality. While AI has improved SP efficiency, existing methods struggle to capture weak early temporal signals. This paper introduces a Multi-Endogenous-view Representation Enhancement (MERE) mechanism to construct enriched feature views, coupled with a Cascaded Dual-convolution Time-series Attention (CDTA) module for multi-scale temporal representation learning. The proposed MEET-Sepsis framework achieves competitive prediction accuracy using only 20\% of the ICU monitoring time required by SOTA methods, significantly advancing early SP. Extensive validation confirms its efficacy. Code is available at: https://github.com/yueliangy/MEET-Sepsis.",

keywords = "Attention mechanism, Classification, Early sepsis prediction, Multi-view learning, Supervised learning, Time-series",

author = "Zexi Tan and Tao Xie and Binbin Sun and Xiang Zhang and Yiqun Zhang and Cheung, \{Yiu Ming\}",

note = "This work was supported in part by the National Natural Science Foundation of China (NSFC) under Grant 62476063, the NSFC/Research Grants Council (RGC) Joint Research Scheme under Grant N\_HKBU214/21, the Natural Science Foundation of Guangdong Province under Grant 2025A1515011293, the General Research Fund of RGC under Grants 12202622 and 12201323, the RGC Senior Research Fellow Scheme under Grant SRFS23242S02, the General Projects of Shenzhen Science and Technology Program under Grant JCYJ202408 13115124032, and the Shenzhen Maternity and Child Healthcare Hospital under Grant FYA2022018. Publisher Copyright: {\textcopyright} The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2026.; 22nd Pacific Rim International Conference on Artificial Intelligence, PRICAI 2025, PRICAI 2025 ; Conference date: 17-11-2025 Through 21-11-2025",

year = "2026",

month = may,

day = "8",

doi = "10.1007/978-981-95-7084-3\_47",

language = "English",

isbn = "9789819570836",

series = "Lecture Notes in Computer Science",

publisher = "Springer",

pages = "678--686",

editor = "Yi Mei and Chao Qian and Quan Bai and Bing Xue and Sankalp Khanna",

booktitle = "PRICAI 2025: Trends in Artificial Intelligence",

address = "Singapore",

url = "https://link.springer.com/book/10.1007/978-981-95-7084-3",

}

Tan, Z, Xie, T, Sun, B, Zhang, X, Zhang, Y & Cheung, YM 2026, MEET-Sepsis: Multi-Endogenous-View Enhanced Time-Series Representation Learning for Early Sepsis Prediction. in Y Mei, C Qian, Q Bai, B Xue & S Khanna (eds), PRICAI 2025: Trends in Artificial Intelligence: 22nd Pacific Rim International Conference on Artificial Intelligence, PRICAI 2025, Wellington, New Zealand, November 17–21, 2025, Proceedings, Part V. Lecture Notes in Computer Science, vol. 16455, Lecture Notes in Artificial Intelligence, PRICAI: Pacific Rim International Conference on Artificial Intelligence, Springer, pp. 678-686, 22nd Pacific Rim International Conference on Artificial Intelligence, PRICAI 2025, Wellington, New Zealand, 17/11/25. https://doi.org/10.1007/978-981-95-7084-3_47

MEET-Sepsis: Multi-Endogenous-View Enhanced Time-Series Representation Learning for Early Sepsis Prediction. / Tan, Zexi (Co-first author); Xie, Tao (Co-first author); Sun, Binbin et al.
PRICAI 2025: Trends in Artificial Intelligence: 22nd Pacific Rim International Conference on Artificial Intelligence, PRICAI 2025, Wellington, New Zealand, November 17–21, 2025, Proceedings, Part V. ed. / Yi Mei; Chao Qian; Quan Bai; Bing Xue; Sankalp Khanna. Springer, 2026. p. 678-686 (Lecture Notes in Computer Science; Vol. 16455), (Lecture Notes in Artificial Intelligence), (PRICAI: Pacific Rim International Conference on Artificial Intelligence).

Research output: Chapter in book/report/conference proceeding › Conference proceeding › peer-review

TY - GEN

T1 - MEET-Sepsis: Multi-Endogenous-View Enhanced Time-Series Representation Learning for Early Sepsis Prediction

AU - Tan, Zexi

AU - Xie, Tao

AU - Sun, Binbin

AU - Zhang, Xiang

AU - Zhang, Yiqun

AU - Cheung, Yiu Ming

N1 - This work was supported in part by the National Natural Science Foundation of China (NSFC) under Grant 62476063, the NSFC/Research Grants Council (RGC) Joint Research Scheme under Grant N_HKBU214/21, the Natural Science Foundation of Guangdong Province under Grant 2025A1515011293, the General Research Fund of RGC under Grants 12202622 and 12201323, the RGC Senior Research Fellow Scheme under Grant SRFS23242S02, the General Projects of Shenzhen Science and Technology Program under Grant JCYJ202408 13115124032, and the Shenzhen Maternity and Child Healthcare Hospital under Grant FYA2022018. Publisher Copyright: © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2026.

PY - 2026/5/8

Y1 - 2026/5/8

N2 - Sepsis is a life-threatening infectious syndrome associated with high mortality in intensive care units (ICUs). Early and accurate sepsis prediction (SP) is critical for timely intervention, yet remains challenging due to subtle early manifestations and rapidly escalating mortality. While AI has improved SP efficiency, existing methods struggle to capture weak early temporal signals. This paper introduces a Multi-Endogenous-view Representation Enhancement (MERE) mechanism to construct enriched feature views, coupled with a Cascaded Dual-convolution Time-series Attention (CDTA) module for multi-scale temporal representation learning. The proposed MEET-Sepsis framework achieves competitive prediction accuracy using only 20% of the ICU monitoring time required by SOTA methods, significantly advancing early SP. Extensive validation confirms its efficacy. Code is available at: https://github.com/yueliangy/MEET-Sepsis.

AB - Sepsis is a life-threatening infectious syndrome associated with high mortality in intensive care units (ICUs). Early and accurate sepsis prediction (SP) is critical for timely intervention, yet remains challenging due to subtle early manifestations and rapidly escalating mortality. While AI has improved SP efficiency, existing methods struggle to capture weak early temporal signals. This paper introduces a Multi-Endogenous-view Representation Enhancement (MERE) mechanism to construct enriched feature views, coupled with a Cascaded Dual-convolution Time-series Attention (CDTA) module for multi-scale temporal representation learning. The proposed MEET-Sepsis framework achieves competitive prediction accuracy using only 20% of the ICU monitoring time required by SOTA methods, significantly advancing early SP. Extensive validation confirms its efficacy. Code is available at: https://github.com/yueliangy/MEET-Sepsis.

KW - Attention mechanism

KW - Classification

KW - Early sepsis prediction

KW - Multi-view learning

KW - Supervised learning

KW - Time-series

UR - https://www.scopus.com/pages/publications/105040392595

U2 - 10.1007/978-981-95-7084-3_47

DO - 10.1007/978-981-95-7084-3_47

M3 - Conference proceeding

AN - SCOPUS:105040392595

SN - 9789819570836

T3 - Lecture Notes in Computer Science

SP - 678

EP - 686

BT - PRICAI 2025: Trends in Artificial Intelligence

A2 - Mei, Yi

A2 - Qian, Chao

A2 - Bai, Quan

A2 - Xue, Bing

A2 - Khanna, Sankalp

PB - Springer

T2 - 22nd Pacific Rim International Conference on Artificial Intelligence, PRICAI 2025

Y2 - 17 November 2025 through 21 November 2025

ER -

Tan Z, Xie T, Sun B, Zhang X, Zhang Y, Cheung YM. MEET-Sepsis: Multi-Endogenous-View Enhanced Time-Series Representation Learning for Early Sepsis Prediction. In Mei Y, Qian C, Bai Q, Xue B, Khanna S, editors, PRICAI 2025: Trends in Artificial Intelligence: 22nd Pacific Rim International Conference on Artificial Intelligence, PRICAI 2025, Wellington, New Zealand, November 17–21, 2025, Proceedings, Part V. Springer. 2026. p. 678-686. (Lecture Notes in Computer Science). (Lecture Notes in Artificial Intelligence). (PRICAI: Pacific Rim International Conference on Artificial Intelligence). doi: 10.1007/978-981-95-7084-3_47