Home > Research > Publications & Outputs > Human-in-the-Loop Learning with LLMs for Effici...

Research

Links

Text available via DOI:

Keywords

View graph of relations

Human-in-the-Loop Learning with LLMs for Efficient RASE Tagging in Building Compliance Regulations

Research output: Contribution to Journal/Magazine › Journal article › peer-review

Published

Standard

Human-in-the-Loop Learning with LLMs for Efficient RASE Tagging in Building Compliance Regulations. / Al-Turki, D.; Hettiarachchi, H.; Gaber, M.M. et al.
In: IEEE Access, Vol. 12, 31.12.2024, p. 185291-185306.

Research output: Contribution to Journal/Magazine › Journal article › peer-review

Harvard

Al-Turki, D, Hettiarachchi, H, Gaber, MM, Abdelsamea, MM, Basurra, S, Iranmanesh, S, Saadany, H & Vakaj, E 2024, 'Human-in-the-Loop Learning with LLMs for Efficient RASE Tagging in Building Compliance Regulations', IEEE Access, vol. 12, pp. 185291-185306. https://doi.org/10.1109/ACCESS.2024.3512434

APA

Al-Turki, D., Hettiarachchi, H., Gaber, M. M., Abdelsamea, M. M., Basurra, S., Iranmanesh, S., Saadany, H., & Vakaj, E. (2024). Human-in-the-Loop Learning with LLMs for Efficient RASE Tagging in Building Compliance Regulations. IEEE Access, 12, 185291-185306. https://doi.org/10.1109/ACCESS.2024.3512434

Vancouver

Al-Turki D, Hettiarachchi H, Gaber MM, Abdelsamea MM, Basurra S, Iranmanesh S et al. Human-in-the-Loop Learning with LLMs for Efficient RASE Tagging in Building Compliance Regulations. IEEE Access. 2024 Dec 31;12:185291-185306. Epub 2024 Dec 5. doi: 10.1109/ACCESS.2024.3512434

Author

Al-Turki, D. ; Hettiarachchi, H. ; Gaber, M.M. et al. / Human-in-the-Loop Learning with LLMs for Efficient RASE Tagging in Building Compliance Regulations. In: IEEE Access. 2024 ; Vol. 12. pp. 185291-185306.

Bibtex

@article{ed72b47cdd904f6ab4f645995b7cb75e,
title = "Human-in-the-Loop Learning with LLMs for Efficient RASE Tagging in Building Compliance Regulations",
abstract = "Automated compliance checking (ACC) in the Architecture, Engineering, and Construction (AEC) sector represents a pivotal task which is traditionally executed manually, demanding significant time and labor. This work investigates the automation of the Requirement, Applicability, Selection, and Exception (RASE) methodology for building regulatory compliance through the utilization of Large Language Models (LLMs) and active learning techniques. Specifically, we focus on the development and assessment of a system using the OpenAI GPT-4o model to transmute building regulation texts into structured YAML formats conducive to ACC processes. The study encompasses three experimental paradigms: few-shot learning, fine-tuning learning, and progressive active learning. Initial results from the few-shot learning experiment illustrate the model's preliminary ability to interpret and process regulatory texts with limited examples. Fine-tuning enhances model performance by training it on a specialized dataset, thereby improving structural and textual accuracy. Progressive active learning, by iteratively incorporating expert feedback, further refines the accuracy of the model. The findings demonstrate substantial enhancements in both structural and semantic accuracies of the generated YAML files, underscoring the potential of integrating LLMs with active learning to streamline regulatory compliance automation. The methodologies and results presented here offer a comprehensive framework for advancing future research and practical applications in the domain of automated regulatory compliance. {\textcopyright} 2013 IEEE.",
keywords = "Active Learning, Automated Compliance Checking, Large Language Models, RASE, Adversarial machine learning, Building codes, Federated learning, Regulatory compliance, Semantics, Zero-shot learning, Architecture engineering, Automated compliance checking, Compliance regulations, Fine tuning, Human-in-the-loop, In-buildings, Language model, Large language model, Requirement, applicability, selection, and exception, Active learning",
author = "D. Al-Turki and H. Hettiarachchi and M.M. Gaber and M.M. Abdelsamea and S. Basurra and S. Iranmanesh and H. Saadany and E. Vakaj",
year = "2024",
month = dec,
day = "31",
doi = "10.1109/ACCESS.2024.3512434",
language = "English",
volume = "12",
pages = "185291--185306",
journal = "IEEE Access",
issn = "2169-3536",
publisher = "Institute of Electrical and Electronics Engineers Inc.",

}

RIS

TY - JOUR

T1 - Human-in-the-Loop Learning with LLMs for Efficient RASE Tagging in Building Compliance Regulations

AU - Al-Turki, D.

AU - Hettiarachchi, H.

AU - Gaber, M.M.

AU - Abdelsamea, M.M.

AU - Basurra, S.

AU - Iranmanesh, S.

AU - Saadany, H.

AU - Vakaj, E.

PY - 2024/12/31

Y1 - 2024/12/31

N2 - Automated compliance checking (ACC) in the Architecture, Engineering, and Construction (AEC) sector represents a pivotal task which is traditionally executed manually, demanding significant time and labor. This work investigates the automation of the Requirement, Applicability, Selection, and Exception (RASE) methodology for building regulatory compliance through the utilization of Large Language Models (LLMs) and active learning techniques. Specifically, we focus on the development and assessment of a system using the OpenAI GPT-4o model to transmute building regulation texts into structured YAML formats conducive to ACC processes. The study encompasses three experimental paradigms: few-shot learning, fine-tuning learning, and progressive active learning. Initial results from the few-shot learning experiment illustrate the model's preliminary ability to interpret and process regulatory texts with limited examples. Fine-tuning enhances model performance by training it on a specialized dataset, thereby improving structural and textual accuracy. Progressive active learning, by iteratively incorporating expert feedback, further refines the accuracy of the model. The findings demonstrate substantial enhancements in both structural and semantic accuracies of the generated YAML files, underscoring the potential of integrating LLMs with active learning to streamline regulatory compliance automation. The methodologies and results presented here offer a comprehensive framework for advancing future research and practical applications in the domain of automated regulatory compliance. © 2013 IEEE.

AB - Automated compliance checking (ACC) in the Architecture, Engineering, and Construction (AEC) sector represents a pivotal task which is traditionally executed manually, demanding significant time and labor. This work investigates the automation of the Requirement, Applicability, Selection, and Exception (RASE) methodology for building regulatory compliance through the utilization of Large Language Models (LLMs) and active learning techniques. Specifically, we focus on the development and assessment of a system using the OpenAI GPT-4o model to transmute building regulation texts into structured YAML formats conducive to ACC processes. The study encompasses three experimental paradigms: few-shot learning, fine-tuning learning, and progressive active learning. Initial results from the few-shot learning experiment illustrate the model's preliminary ability to interpret and process regulatory texts with limited examples. Fine-tuning enhances model performance by training it on a specialized dataset, thereby improving structural and textual accuracy. Progressive active learning, by iteratively incorporating expert feedback, further refines the accuracy of the model. The findings demonstrate substantial enhancements in both structural and semantic accuracies of the generated YAML files, underscoring the potential of integrating LLMs with active learning to streamline regulatory compliance automation. The methodologies and results presented here offer a comprehensive framework for advancing future research and practical applications in the domain of automated regulatory compliance. © 2013 IEEE.

KW - Active Learning

KW - Automated Compliance Checking

KW - Large Language Models

KW - RASE

KW - Adversarial machine learning

KW - Building codes

KW - Federated learning

KW - Regulatory compliance

KW - Semantics

KW - Zero-shot learning

KW - Architecture engineering

KW - Automated compliance checking

KW - Compliance regulations

KW - Fine tuning

KW - Human-in-the-loop

KW - In-buildings

KW - Language model

KW - Large language model

KW - Requirement, applicability, selection, and exception

KW - Active learning

U2 - 10.1109/ACCESS.2024.3512434

DO - 10.1109/ACCESS.2024.3512434

M3 - Journal article

VL - 12

SP - 185291

EP - 185306

JO - IEEE Access

JF - IEEE Access

SN - 2169-3536

ER -