Knowledge Extraction from Documents of Legal content

About

The project “KnEDLe - Knowledge Extraction from Documents of Legal content” is a partnership among FAPDF (Fundação de Apoio à Pesquisa do Distrito Federal), UnB (the University of Brasília) and Finatec (Fundação de Empreendimentos Científicos e Tecnológicos), sponsored by FAPDF. This project was proposed in order to employ official publications as a research object and to extract knowledge. The objective is to develop intelligent tools for extracting structured information from such publications, aiming to facilitate the search and retrieval of information, increasing government transparency and facilitating audit tasks and detecting problems related to the use of public resources.

Official publications such as the Diário Oficial do Distrito Federal (DODF) are sources of information on all official government acts. Although these documents are rich in knowledge, analysing these texts manually by specialists is a complex and unfeasible task considering the growing volume of documents, the result of the frequent number of publications in the Distrito Federal Government's (GDF) communication vehicle.

This scenario is appropriate to employ computational techniques based on text mining and information visualization, in order to discover implicit and relevant knowledge in large textual data sets. It is known that these computational techniques receive data in a structured format. However, as DODF editions are originally published in unstructured format and in natural language, it is required to use techniques to prepare strategies in order to make the necessary adaptations to apply.

[FAPDF Post] Artificial Intelligence in the extraction of information in the Official Gazette of the DF

Events

VI Workshop
4 2023
This event will take place on March 24, 2023, in the TCDF auditorium - Federal District Court of Auditors. - Palácio Costa e Silva, Praça do Buriti 70075-901 - Brasília, DF
- 14h00 - Opening - Thiago Faleiros
- 14h10 - The Knedle Project - Vinícius Borges, Luis Paulo Faina and Andrei Queiroz
- Annotation, NitoTat and VisNote
- DODFMiner and Knedash
- 14h30 - DODFMiner CLI - Gabriel Guimarães
- 14h50 - DODFMiner Lib - Bruno Esteves
- 15h10 - Knedash and Timeline - Ian F. P. Ferreira
- 15h30 - Discussions and Closing
V Workshop
8 2022
This event will take place on July 29, 2022, and the meeting can be accessed via the link above.
Subscription for KnEDLe Project V Workshop. Subscription is required to obtain the certificate issued by the TCDF
Subscribe
- 14h00 - Opening - Thiago Faleiros
- 14h10 - Knedash - Jonatas G. B. da Silva
- 14h30 - Label Studio - Gabriel Nogueira and Vitor Vasconcelos
- 14h50 - DODFKge - Rafael A. Soares
- 15h10 - Bids and Contracts - Gabriel M. C. Guimarães and Ian F. P. Ferreira
- 15h30 - Discussions and Closing
IV Workshop
3 October 2022
This event took place in February 2022, and its video recording is available from the link above.
- 14h00 - Opening - Thiago Faleiros
- 14h10 - DODFMiner - Vitor Araruna
- 14h30 - DODFKge - Rafael Amaral
- 14h50 - Bids and Contracts Timeline - Daniel de Sousa
- 15h20 - Closing
III Workshop
9 May 2021
This event took place in August 2021, and its video recording is available from the link above.
- 14h00 - Opening - Teo de Campos and Thiago Faleiros
- 14h10 - DODFMiner - Thiago Faleiros
- 14h30 - Deep Active-Self learning applied to NER - José Reinaldo
- 14h50 - Annotation of Personnel Acts - Vinícius Borges
- 15h10 - Visnote - Tatiana Franco Pereira
- 15h30 - Bids and Contracts Timeline - Nilton Correia
- 15h50 - Next Steps Planning
- 16h00 - Closing
II Workshop
3 October 2021
This event took place in February 2021, and its video recording is available from the link above.
- 14h00 - Opening - Teófilo de Campos
- 14h10 - DodfMiner - Khalil Carsten
- 14h40 - NidoTat - Lívia Fonseca e Matheus Stauffer
- 15h10 - VisNote - Tatiana Franco
- 15h25 - Paper summary accepted on JURIX - Pedro Luz
- 15h40 - Next Steps
- 16h00 - Closing
I Workshop
8 2020
This event took place in July 2020, and its video recording is available from the link above.
- 14h00 - Opening - Gladston da Silva e Teófilo de Campos
- 14h15 - Publication at PROPOR 2020 - Pedro Luz
- 14h45 - ToolBox - Thiago de Paulo, Khalil Carsten e Renato Nobre
- 15h30 - Strategies for DODF text annotation - Vinicius Borges
- 15h45 - Round Table
- 16h00 - Closing

Publications

LayoutQT—Layout Quadrant Tags to embed visual features for document analysis
• Patricia Medyna Lauritzen de Lucena Drumond • Lindeberg Pessoa Leite • Fabricio A. Braz • Teofilo E. de Campos
Sequence-aware multimodal page classification of Brazilian legal documents
• Pedro H. Luz de Araujo • Ana Paula G. S. de Almeida • Fabricio A. Braz • Nilton C. da Silva • Flavio de Barros Vidal • Teofilo E. de Campos
Aprendizado de maquina para a sugestão automática de correções em anotações de dados textuais
• Tatiana F. Pereira • Vinícius R. P. Borges

See all publications

Publications

Articles

LayoutQT—Layout Quadrant Tags to embed visual features for document analysis
• Patricia Medyna Lauritzen de Lucena Drumond • Lindeberg Pessoa Leite • Fabricio A. Braz • Teofilo E. de Campos
Sequence-aware multimodal page classification of Brazilian legal documents
• Pedro H. Luz de Araujo • Ana Paula G. S. de Almeida • Fabricio A. Braz • Nilton C. da Silva • Flavio de Barros Vidal • Teofilo E. de Campos
Named Entity Recognition Approaches Applied to Legal Document Segmentation.
SILVA, F. ; GUIMARAES, G. ; QUEIROZ, A. ; BORGES, V. R. P. ; FALEIROS, THIAGO ; GARCIA, L. ; MARCACIN, R. M.
In: Symposium on Knowledge Discovery, Mining and Learning -- KDMILE 22, 2022.
On the use of Query by Committee for Human-in-the-Loop Named Entity Recognition.
NOGUEIRA, G. S.; OLIVEIRA, V. V.; MARCACINI, R.
In: Symposium on Knowledge Discovery, Mining and Learning -- KDMILE 22, 2022.
Data Augmented 3D Semantic Scene Completion With 2D Segmentation Priors
• Aloisio Dourado • Frederico Guth • Teófilo E. de Campos
IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), 2022
Learn by Guessing: Multi-step Pseudo-label Refinement for Person Re-Identification
• Tiago de C. G. Pereira • Teófilo E. de Campos
17th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Aplications
Topic Modelling Brazilian Supreme Court Lawsuits
• Pedro H. Luz de Araujo • Teófilo E. de Campos
33rd International Conference on Legal Knowledge and Information Systems (JURIX 2020)
VisNote: A Tool For Interactive Visual Exploration And Text Annotation Of Government Gazettes
• Tatiana F. Pereira • Lívia G. C Fonseca • Matheus S. V. de Oliveira • Teofilo E. Campos • Vinícius R. P. Borges
33rd International Conference on Graphics, Patterns and Images (SIBGRAPI 2020)
Topic Modelling Brazilian Supreme Court Lawsuits
• Pedro H. Luz de Araujo • Teófilo E. de Campos
33rd International Conference on Legal Knowledge and Information Systems (JURIX 2020)
Inferring The Source Of Official Texts: Can SVM Beat ULMFiT?
• Pedro H. Luz de Araujo • Teófilo E. de Campos • Marcelo M. Silva de Sousa
PROPOR 2020 - International Conference on the Computational Processing of Portuguese
Victor: A Dataset For Brazilian Legal Documents Classification
• Pedro Henrique Luz de Araujo • Teófilo Emídio de Campos • Fabricio Ataides Braz • Nilton Correia da Silva
Language Resources and Evaluation Conference (LREC)

Tools

DODFMiner

Data Extraction Library

Extraction of data from documents in PDF format referring to the,publications of the Official Gazette of the Federal District.

Documentation Repository

Knedash

The Official Gazette of the Federal District in a simple way

Through AI, Knedash performs queries and extracts acts and entities from the Official Gazette for the Federal District Court of Auditors.

Knedash Documentation Repository

LabelStudio

Open-source Human-in-the-loop (HITL) data annotation tool for named entity recognition (NER) tasks

Our tool is publicly available and can be integrated with pre-trained language models to speed up the labeling task via active learning, transfer learning and fine-tuning. We evaluated the tool considering the legal domain through the Query by Committee (QBC) method published in the article link below

LabelStudio Reference Article Repository

VisNote

A visualization-based tool developed for exploring and annotating textual data from Government Gazettes.

A visualization tool for publications of the Official Gazette of the Federal District that mainly employs visualizations based on the positioning of points, such as t-Distributed Stochastic Neighbor Embedding (t-SNE) and Uniform Manifold Approximation and Projection (UMAP). The idea is to involve the specialist in knowledge discovery tasks in official publications with the support of visualizations that graphically represent the main implicit patterns in the associated texts.

Repository

NidoTat

A custom annotation tool from the TeamTat tool for the purposes of the KnEDLe project.

This tool is useful for creating corpus for named entity recognition tasks, with support for collaborative annotation (multiple annotators) and centered on the person responsible for conducting the annotation process.

Repository