The presentation at MCIS Corfu (look for the title in AIS library, for the full paper). "EU-Wide Legal Text Mining using Big Data Processsing Infrastructures"

1. EU-WIDE LEGAL TEXT MINING
USING BIG DATA PROCESSING
INFRASTRUCTURE
Assoc. Professor Yannis Charalabidis
University of the Aegean
yannisx@aegean.gr
1

2. Overview
• Purpose
• Research Methodology
• Theoretical Background
• Workshops Findings
• A Framework for Legal Text Mining using Big Data
• The proposed ICT Architecture
• Conclusions
• Further Research
2

3. Purpose
Address the challenge of fragmented information in
the legal domain:
• Integration of automated translation services and
utilization of HPC resources
• Mining national legal data portals and EU sources
of legal information
• Aggregating the retrieved information with the
data sustained public administrations and law-
making bodies
We estimate that all legal artefacts databases will
contain more than 1 trillion words in 21 different
languages, corresponding to about 10 million
“volumes” of classical books, when another 5,000
such “volumes” will be added for study, on a daily
basis.
3

4. Research Methodology
4
Literature Review –
Identify established
methods and tools
1
Review of existing
infrastructures
2
Two workshops to
identify the need
(Hellenic Parliament
and Austrian
Parliament)
3
Framework for Legal
Text Mining using Big
Data
4

5. Theoretical Background
• Literature Review
• Legal Informatics refer to the application of Information Technology within the context of legal environment (Erdelez et al,
1997)
• Defined by Sartor and Francesconi (2010) as the «theory and practice of computable law, i.e. the cooperation between
humans and machines in legal problem-solving»
• The standard text analysis pipeline performs several levels of analysis: morphological, syntactic, semantic, and discourse
(Lacity & Janson, 1994)
• Review of existing infrastructure
• Peri Nomou System has the ability to extract data from laws such as correlations and vocabulary and to modify structural
elements of Greek Laws
• Nomos: Legislative database services - Providing the entire Greek legislation
• Openlaws.eu: European Project – A network of legislation, case law, legal literature and legal expert which automatically
collects data from different sources
5

6. MANYLAWS Services
The planed services are:
• Parallel search in many EU member-state legal frameworks (through parallel translation of search terms)
• Analysis of the embodiment/transposition of an EU Directive in a National Legal Framework
• Analysis of references to the European Legislation by National Laws
• Comparative analysis of two relevant laws from different EU member states
• Comparative analysis of connected laws from the same member state
• Timeline analysis for all legal elements
• Interrelation of laws and news or social media posts, including sentiment analysis
• Various geo-related visualisations (e.g. EU maps indicating different parameters)
• Various text-related visualisations (e.g. wordle, sentiment graphs, interrelation maps, etc) and other common
visual aids (e.g. graphs, charts, tables, etc) 6

7. A Framework for Legal Text Mining using Big
Data
• Data
• All the legal artefacts (European Parliament, European Commission, EU Council, National
Parliaments)
• Processing
• The Pre-processing stage is responsible to prepare data for text mining
• The data are converted to structured data using text mining techniques in order to offer a variety of
services
• Services
• Services are to be provided towards citizens, businesses and administrations, based on the most
common needs of each user type
7

8. MANYLAWS
The ICT
Architecture
8

9. Conclusions
• We have proposed a framework analysis for legal text mining using big data
that facilitates to solve legal issues
• This framework helps citizens and enterprises to follow the latest legislation
• The establishment of the proposed framework foster transparency and
promote social inclusion in lawmaking
• Citizens find and understand relevant legislation, case law, and other legal
information and so that they understand how EU-law is created and
implemented
9

10. Further steps
10
Implementation of the proposed framework in steps: 4 cores (working) - 40 cores (6
months) - 40.000 cores (18 months) - 400.000 cores (24m - if needed)
Develop the first set of services / scenarios at MVP level (6 m)
Validation by public admin, legal, ICT experts, professionals and citizens. Join the
team, to help us build our next, bigger project covering the globe (from 2, to 28, to !