Combining rule-based and machine learning methods for efficient information extraction from enforcement decisions

Harry Nan; Maarten Marx; Johan Wolswinkel

doi:10.3233/FAIA241262

Combining rule-based and machine learning methods for efficient information extraction from enforcement decisions

Harry Nan, Maarten Marx, Johan Wolswinkel

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

Abstract

This paper presents an effective and efficient approach for automatic extraction of key features from enforcement decisions, such as their legal basis and their legal effect, by strategically applying a Large Language Model (LLM) on top of rule-based methods. Initially, rule-based methods identify candidate sentences within these decisions containing these features, after which these sentences are analyzed by GPT-3.5 to extract the features. This approach is efficient as it reduces the input and number of resources needed for effective and context aware information extraction. Furthermore, other features that have not been subject to a rule-based selection first can be extracted by an LLM from the same set of candidate sentences when they exist in close proximity of each other.

Original language	English
Title of host publication	JURIX 2024
Subtitle of host publication	The thirty-seventh annual conference, Brno, Czech Republic, 11-13 December 2024
Editors	Jaromir Savelka, Jakub Harasta, Tereza Novotna, Jakub Misek
Publisher	IOS Press
Pages	321-326
Number of pages	6
Volume	395
ISBN (Electronic)	9781643685625
DOIs	https://doi.org/10.3233/FAIA241262
Publication status	Published - 5 Dec 2024
Event	37th International Conference on Legal Knowledge and Information Systems - Faculty of Law, Masaryk University, Brno, Czech Republic Duration: 11 Dec 2024 → 13 Dec 2024 https://jurix2024.law.muni.cz/

Publication series

Name	Frontiers in Artificial Intelligence and Applications

Conference

Conference	37th International Conference on Legal Knowledge and Information Systems
Country/Territory	Czech Republic
City	Brno
Period	11/12/24 → 13/12/24
Internet address	https://jurix2024.law.muni.cz/

Keywords

information extraction
enforcement decisions
rule-based methods
machine learning methods
large language model
text segmentation

Access to Document

10.3233/FAIA241262Licence: CC BY-NC

Cite this

Nan, H., Marx, M., & Wolswinkel, J. (2024). Combining rule-based and machine learning methods for efficient information extraction from enforcement decisions. In J. Savelka, J. Harasta, T. Novotna, & J. Misek (Eds.), JURIX 2024: The thirty-seventh annual conference, Brno, Czech Republic, 11-13 December 2024 (Vol. 395, pp. 321-326). (Frontiers in Artificial Intelligence and Applications). IOS Press. https://doi.org/10.3233/FAIA241262

Nan, Harry ; Marx, Maarten ; Wolswinkel, Johan. / Combining rule-based and machine learning methods for efficient information extraction from enforcement decisions. JURIX 2024: The thirty-seventh annual conference, Brno, Czech Republic, 11-13 December 2024. editor / Jaromir Savelka ; Jakub Harasta ; Tereza Novotna ; Jakub Misek. Vol. 395 IOS Press, 2024. pp. 321-326 (Frontiers in Artificial Intelligence and Applications).

@inbook{c31d97e448f24a8583d49d63902065b9,

title = "Combining rule-based and machine learning methods for efficient information extraction from enforcement decisions",

abstract = "This paper presents an effective and efficient approach for automatic extraction of key features from enforcement decisions, such as their legal basis and their legal effect, by strategically applying a Large Language Model (LLM) on top of rule-based methods. Initially, rule-based methods identify candidate sentences within these decisions containing these features, after which these sentences are analyzed by GPT-3.5 to extract the features. This approach is efficient as it reduces the input and number of resources needed for effective and context aware information extraction. Furthermore, other features that have not been subject to a rule-based selection first can be extracted by an LLM from the same set of candidate sentences when they exist in close proximity of each other.",

keywords = "information extraction, enforcement decisions, rule-based methods, machine learning methods, large language model, text segmentation",

author = "Harry Nan and Maarten Marx and Johan Wolswinkel",

year = "2024",

month = dec,

day = "5",

doi = "10.3233/FAIA241262",

language = "English",

volume = "395",

series = "Frontiers in Artificial Intelligence and Applications",

publisher = "IOS Press",

pages = "321--326",

editor = "Jaromir Savelka and Harasta, { Jakub} and Tereza Novotna and { Misek}, { Jakub }",

booktitle = "JURIX 2024",

address = "Netherlands",

note = "37th International Conference on Legal Knowledge and Information Systems ; Conference date: 11-12-2024 Through 13-12-2024",

url = "https://jurix2024.law.muni.cz/",

}

Nan, H, Marx, M & Wolswinkel, J 2024, Combining rule-based and machine learning methods for efficient information extraction from enforcement decisions. in J Savelka, J Harasta, T Novotna & J Misek (eds), JURIX 2024: The thirty-seventh annual conference, Brno, Czech Republic, 11-13 December 2024. vol. 395, Frontiers in Artificial Intelligence and Applications, IOS Press, pp. 321-326, 37th International Conference on Legal Knowledge and Information Systems, Brno, Czech Republic, 11/12/24. https://doi.org/10.3233/FAIA241262

Combining rule-based and machine learning methods for efficient information extraction from enforcement decisions. / Nan, Harry; Marx, Maarten; Wolswinkel, Johan.
JURIX 2024: The thirty-seventh annual conference, Brno, Czech Republic, 11-13 December 2024. ed. / Jaromir Savelka; Jakub Harasta; Tereza Novotna; Jakub Misek. Vol. 395 IOS Press, 2024. p. 321-326 (Frontiers in Artificial Intelligence and Applications).

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

TY - CHAP

T1 - Combining rule-based and machine learning methods for efficient information extraction from enforcement decisions

AU - Nan, Harry

AU - Marx, Maarten

AU - Wolswinkel, Johan

PY - 2024/12/5

Y1 - 2024/12/5

N2 - This paper presents an effective and efficient approach for automatic extraction of key features from enforcement decisions, such as their legal basis and their legal effect, by strategically applying a Large Language Model (LLM) on top of rule-based methods. Initially, rule-based methods identify candidate sentences within these decisions containing these features, after which these sentences are analyzed by GPT-3.5 to extract the features. This approach is efficient as it reduces the input and number of resources needed for effective and context aware information extraction. Furthermore, other features that have not been subject to a rule-based selection first can be extracted by an LLM from the same set of candidate sentences when they exist in close proximity of each other.

AB - This paper presents an effective and efficient approach for automatic extraction of key features from enforcement decisions, such as their legal basis and their legal effect, by strategically applying a Large Language Model (LLM) on top of rule-based methods. Initially, rule-based methods identify candidate sentences within these decisions containing these features, after which these sentences are analyzed by GPT-3.5 to extract the features. This approach is efficient as it reduces the input and number of resources needed for effective and context aware information extraction. Furthermore, other features that have not been subject to a rule-based selection first can be extracted by an LLM from the same set of candidate sentences when they exist in close proximity of each other.

KW - information extraction

KW - enforcement decisions

KW - rule-based methods

KW - machine learning methods

KW - large language model

KW - text segmentation

U2 - 10.3233/FAIA241262

DO - 10.3233/FAIA241262

M3 - Chapter

VL - 395

T3 - Frontiers in Artificial Intelligence and Applications

SP - 321

EP - 326

BT - JURIX 2024

A2 - Savelka, Jaromir

A2 - Harasta, Jakub

A2 - Novotna, Tereza

A2 - Misek, Jakub

PB - IOS Press

T2 - 37th International Conference on Legal Knowledge and Information Systems

Y2 - 11 December 2024 through 13 December 2024

ER -

Nan H, Marx M, Wolswinkel J. Combining rule-based and machine learning methods for efficient information extraction from enforcement decisions. In Savelka J, Harasta J, Novotna T, Misek J, editors, JURIX 2024: The thirty-seventh annual conference, Brno, Czech Republic, 11-13 December 2024. Vol. 395. IOS Press. 2024. p. 321-326. (Frontiers in Artificial Intelligence and Applications). doi: 10.3233/FAIA241262

Combining rule-based and machine learning methods for efficient information extraction from enforcement decisions

Abstract

Publication series

Conference

Keywords

Access to Document

Fingerprint

Cite this