Automated Retrieval of Metadata from Legal Texts

SCHEME: Proof-of-Concept

CALL: 2017

DOMAIN: IS - Information and Communication Technologies


LAST NAME: Sabetzadeh



HOST INSTITUTION: University of Luxembourg

KEYWORDS: Legal Informatics, Legal Metadata, Compliance, Automation, Natural Language Processing

START: 2017-01-01

END: 2018-12-31


Submitted Abstract

Legal informatics is a rapidly emerging topic. The ultimate goal of legal informatics is to improve understanding, dissemination, and compliance with legal texts, notably laws and regulations. A key tenet of legal informatics is leveraging information technologies for making explicit the structure and semantic properties of legal texts. These properties constitute the metadata that needs to be recorded alongside the natural-language content of legal texts, both to facilitate the interpretation of legal texts by humans and to enable advanced analyses such as automated compliance verification and monitoring. Given the sheer scale of legal corpora amassed over decades and centuries, a fully manual creation of legal metadata is extremely laborious and requires a tremendous amount of resources.In the past three years, we have been conducting applied research on automated extraction of metadata from legal texts. This metadata covers structural information, for example, the hierarchical organisation of legal texts and the cross references (citations) in them, as well as semantic information, for example, rights, permissions, obligations, conditions and consequences. As part of this endeavour, we have developed an early prototype based on Natural Language Processing (NLP). Given a collection of legal documents in plain-text format, this prototype automatically processes the documents and extracts metadata from them.The goals of this Proof of Concept (PoC) proposal are two-fold:1. Drawing on the know-how and the lessons learned from developing our early prototype, we will develop a professional and functional toolbox, named ARMLET, for legal metadata extraction.2. We will conduct a thorough market analysis in order to define clear business cases for the toolbox, and to align its development with market needs. We will approach market needs from the perspectives of both governmental and private-sector market players. Both perspectives are critical with ample potential for valorisation.The team involved in ARMLET has already gained substantial traction in the subject matter of the proposal from a close and on-going collaboration with Service Central de Législation (SCL), which is an agency of the Government of Luxembourg. This collaboration significantly boosts the odds of success. In particular, SCL has confirmed its interest in using ARMLET and becoming a first pilot client. SCL has further shown interest in promoting ARMLET in a European arena and helping with exploitation beyond Luxembourg.With respect to the private sector, our preliminary analysis suggests that the valorisation potential for the proposed toolbox is high, provided that the toolbox can be tailored to the needs of specific corporate entities or legal jurisdictions. Understanding and implementing the minimum required level of tailoring will be an essential backbone of this project.The end-product will be implemented as a customisable toolbox with the option to be deployed as a Software-as-a-Service solution. The exact valorisation model will be decided upon subject to the outcome of the PoC and feedback obtained from the targeted market segments. An important aspect of our market analysis during the PoC is to determine the extent to which ARMLET will need to be procured alongside consultancy and training services. Developing a viable consultancy and training strategy is therefore likely to be a part of the proposal.Ultimately and over the course of the PoC, we will pin down the most suitable exploitation route, namely licencing versus spin-off, for commercialising the toolbox and offering, if necessary, consultancy and training services around it.

This site uses cookies. By continuing to use this site, you agree to the use of cookies for analytics purposes. Find out more in our Privacy Statement