Adaptive LCI Data Completion: Integrating Neural Processes and Active Learning for Enhanced Life Cycle Assessment

Wei W. Xing; Hong Chen; Zidong Chen; Zhishan Quan; Bertrand Laratte; Mark Walsh; Jing Pu; Jose L. Casamayor

doi:10.1016/j.procir.2025.01.046

Adaptive LCI Data Completion: Integrating Neural Processes and Active Learning for Enhanced Life Cycle Assessment

Wei W. Xing (Co-first Author)
, Hong Chen (Co-first Author)
, Zidong Chen
, Zhishan Quan
, Bertrand Laratte
, Mark Walsh
, Jing Pu
, Jose L. Casamayor^*

^*Corresponding author for this work

Research output: Journal Publications and Reviews › RGC 21 - Publication in refereed journal › peer-review

1 Citation (Scopus)

Abstract

Accurate and comprehensive Life Cycle Inventory (LCI) data underpins the reliability and accuracy of Life Cycle Assessment (LCA) results. However, LCI data is often incomplete due to data unavailability, which affects the reliability and accuracy of LCA results. To address this issue, this paper introduces a novel approach for LCI data completion based on Neural Processes (NPs) combined with active learning for efficient adaptive refinement of LCI data completion. Experimental results demonstrate that the proposed approach outperforms the state-of-The-Art XGBoost-based method significantly, achieving up to 99% improvement in prediction accuracy. This means that by reducing data requirements by approximately 50% whilst improving predictive accuracy, the proposed AI model can provide more reliable LCA results in less time. © 2025 Elsevier B.V.. All rights reserved.

Original language	English
Pages (from-to)	136-141
Journal	Procedia CIRP
Volume	135
Online published	17 Jul 2025
DOIs	https://doi.org/10.1016/j.procir.2025.01.046
Publication status	Published - 2025
Externally published	Yes
Event	32nd CIRP Conference on Life Cycle Engineering, LCE 2025 - Manchester, United Kingdom Duration: 7 Apr 2025 → 9 Apr 2025

Funding

Thea uthorsw ouldl iket ot hankU KRI (EPSRC-IAA: Grant1 87882 ) and Seco Tools UK Ltd.f ort heir financial support; andR ebecca Holbacha ndD rB inC henf ort heir input and support int hisr esearch.

UN SDGs

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

SDG 9 Industry, Innovation, and Infrastructure
SDG 12 Responsible Consumption and Production

Research Keywords

Active Learning
Life Cycle Assessment
Life Cycle Inventory
Neural Processes

Publisher's Copyright Statement

This full text is made available under CC-BY-NC-ND 4.0. https://creativecommons.org/licenses/by-nc-nd/4.0/

Access Output

10.1016/j.procir.2025.01.046

362377920.pdfFinal Published version, 819 KBLicence: CC BY-NC-ND

Other Access Options

Check CityUHK Library

Permanent Link

Right click to copy link

Cite this

@article{4ad371566d234958a4bd2fb6148f2c0a,

title = "Adaptive LCI Data Completion: Integrating Neural Processes and Active Learning for Enhanced Life Cycle Assessment",

abstract = "Accurate and comprehensive Life Cycle Inventory (LCI) data underpins the reliability and accuracy of Life Cycle Assessment (LCA) results. However, LCI data is often incomplete due to data unavailability, which affects the reliability and accuracy of LCA results. To address this issue, this paper introduces a novel approach for LCI data completion based on Neural Processes (NPs) combined with active learning for efficient adaptive refinement of LCI data completion. Experimental results demonstrate that the proposed approach outperforms the state-of-The-Art XGBoost-based method significantly, achieving up to 99\% improvement in prediction accuracy. This means that by reducing data requirements by approximately 50\% whilst improving predictive accuracy, the proposed AI model can provide more reliable LCA results in less time. {\textcopyright} 2025 Elsevier B.V.. All rights reserved.",

keywords = "Active Learning, Life Cycle Assessment, Life Cycle Inventory, Neural Processes",

author = "Xing, \{Wei W.\} and Hong Chen and Zidong Chen and Zhishan Quan and Bertrand Laratte and Mark Walsh and Jing Pu and Casamayor, \{Jose L.\}",

year = "2025",

doi = "10.1016/j.procir.2025.01.046",

language = "English",

volume = "135",

pages = "136--141",

journal = "Procedia CIRP",

issn = "2212-8271",

publisher = "Elsevier B.V.",

note = "32nd CIRP Conference on Life Cycle Engineering, LCE 2025 ; Conference date: 07-04-2025 Through 09-04-2025",

}

TY - JOUR

T1 - Adaptive LCI Data Completion

T2 - 32nd CIRP Conference on Life Cycle Engineering, LCE 2025

AU - Xing, Wei W.

AU - Chen, Hong

AU - Chen, Zidong

AU - Quan, Zhishan

AU - Laratte, Bertrand

AU - Walsh, Mark

AU - Pu, Jing

AU - Casamayor, Jose L.

PY - 2025

Y1 - 2025

N2 - Accurate and comprehensive Life Cycle Inventory (LCI) data underpins the reliability and accuracy of Life Cycle Assessment (LCA) results. However, LCI data is often incomplete due to data unavailability, which affects the reliability and accuracy of LCA results. To address this issue, this paper introduces a novel approach for LCI data completion based on Neural Processes (NPs) combined with active learning for efficient adaptive refinement of LCI data completion. Experimental results demonstrate that the proposed approach outperforms the state-of-The-Art XGBoost-based method significantly, achieving up to 99% improvement in prediction accuracy. This means that by reducing data requirements by approximately 50% whilst improving predictive accuracy, the proposed AI model can provide more reliable LCA results in less time. © 2025 Elsevier B.V.. All rights reserved.

AB - Accurate and comprehensive Life Cycle Inventory (LCI) data underpins the reliability and accuracy of Life Cycle Assessment (LCA) results. However, LCI data is often incomplete due to data unavailability, which affects the reliability and accuracy of LCA results. To address this issue, this paper introduces a novel approach for LCI data completion based on Neural Processes (NPs) combined with active learning for efficient adaptive refinement of LCI data completion. Experimental results demonstrate that the proposed approach outperforms the state-of-The-Art XGBoost-based method significantly, achieving up to 99% improvement in prediction accuracy. This means that by reducing data requirements by approximately 50% whilst improving predictive accuracy, the proposed AI model can provide more reliable LCA results in less time. © 2025 Elsevier B.V.. All rights reserved.

KW - Active Learning

KW - Life Cycle Assessment

KW - Life Cycle Inventory

KW - Neural Processes

UR - https://www.scopus.com/pages/publications/105011590156

UR - https://www.scopus.com/record/pubmetrics.uri?eid=2-s2.0-105011590156&origin=recordpage

U2 - 10.1016/j.procir.2025.01.046

DO - 10.1016/j.procir.2025.01.046

M3 - RGC 21 - Publication in refereed journal

SN - 2212-8271

VL - 135

SP - 136

EP - 141

JO - Procedia CIRP

JF - Procedia CIRP

Y2 - 7 April 2025 through 9 April 2025

ER -

Adaptive LCI Data Completion: Integrating Neural Processes and Active Learning for Enhanced Life Cycle Assessment

Abstract

Funding

UN SDGs

Research Keywords

Publisher's Copyright Statement

Access Output

Other Access Options

Permanent Link

Other Files and Links

Fingerprint

Cite this