DOING : Intelligent Data – From Data to Knowledge WORKSHOP in ADBIS, TPDL & EDA 2020 joint conferences
- Mírian Halfeld Ferrari – Université d’Orléans, INSA CVL, LIFO EA, France
- Carmem S. Hara – Universidade Federal do Paraná, Curitiba, Brazil
A word about DOING
DOING workshop is connected to the working groups
PROGRAM (first version) – Tuesday, August 25, 2020
9:30-10:00: KEYNOTE: Knowledge Graph Completion and Enrichment in OntoSides using Text Mining. Marie-Christine Rousset (Professor, Laboratoire d’Informatique de Grenoble, senior member of Institut Universitaire de France)
Abstract: Knowledge graph completion and enrichment have become problems of increasing interest for which several supervised and unsupervised techniques have been investigated. The completion and enrichment problems that we consider in this paper target relations of interest guided by the needs in data analytics of domain experts. Our methodology relies on exploiting textual information found in knowledge graphs and consists in experimentally choosing the most appropriate text models and text mining techniques to achieve high precision which is a strong requirement for accurate data analytics. This methodology is illustrated and evaluated on OntoSIDES which is a big knowledge graph at the core of a learning management system used in medical studies in France.
Session 1: NLP for Information Extraction
- 10:15-10:30 Erwan Marchand, Michel Gagnon and Amal Zouaq. Extraction of a Knowledge Graph from French Cultural Heritage Documents
- 10:30-10:45 Joshua Amavi, Mirian Halfeld-Ferrari and Nicolas Hiot. Natural Language Querying System through Entity Enrichment
- 10:45-11:00 Arturo Oncevay, Marco Sobrevilla, Hugo Alatrista-Salas and Andres Melgar. Public Riots in Twitter: Domain-Based Event Filtering during Civil Unrest
- 11:00-11:15 Dimmy Magalhães and Aurora Pozo. Classification of Relationship in Argumentation using Graph Convolutional Network
Session 2: Intelligent Data Management
- 11:30-11:45 Ciro M. Medeiros, Umberto S. Costa, Semyon V. Grigorev and Martin A. Musicante. Recursive Expressions for SPARQL Property Paths
- 11:45-12:00 Guilherme M. Rocha, Piero L. Capelo and Cristina Dutra De Aguiar Ciferri. Healthcare decision-making over a geographic, socioeconomic, and image data warehouse
- 12:00-12:15 Jian Lin and Dongming Xie. OMProv: Provenance Mechanism for Objects in Deep Learning
- 12:15-12:30 Dickson Owuor, Anne Laurent and Joseph Orero. Exploiting IoT data crossings for gradual pattern mining through parallel processing
- 12:30-12:40 Damien Alvarez de Toledo, Laurent D’Orazio, Frederic Andres and Maria Leite. Cooking related Carbon Footprint Evaluation and Optimisation
Aims and scope.
Text are important sources of information and communication in diverse domains. The intelligent, efficient and secure use of this information requires, in most cases, the transformation of unstructured textual data into data sets with some structure, and organized according to an appropriate schema that follows the semantics of an application domain. Indeed, solving the problems of modern society requires interdisciplinary research and information cross-referencing, thus surpassing the simple provision of unstructured data. There is a need for representations that are more flexible, subtle and context-sensitive, which can also be easily accessible via consultation tools and evolve according to these principles. In this context, consultation requires robust and efficient processing of requests, which may involve information analysis, with quality, consistency, and privacy preservation guarantees. Knowledge bases can be built as these new generation infrastructures which support data science queries on a user-friendly framework and are capable of providing the required machinery for advised decision-making.
The workshop focuses on transforming data into information and then into knowledge. The idea is to gather researchers in NLP (Natural Language Processing), DB (Databases), and AI (Artificial Intelligence) to discuss two main problems :
- how to extract information from textual data and represent it in knowledge bases;
- how to propose intelligent methods for handling and maintaining these databases with new forms of requests, including efficient, flexible, and secure analysis mechanisms, adapted to the user, and with quality and privacy preservation guarantees.
This workshop focuses on all aspects concerning these modern infrastructures, giving particular attention (but not limited to) to data related to health and environmental domains.
Topics of interest.
We invite the submission of work-in-progress research addressing various aspects of information extraction from textual data, intelligent and efficient interrogation, and maintenance of knowledge bases. The workshop welcomes submissions of theoretical, technical, experimental, methodological papers, application papers, position papers and papers on experience reports addressing – though not limited to – the following topics:
- Artificial intelligence in databases and information systems
- Data curation, annotation, and provenance
- Data management and analytics
- Data mining and knowledge discovery
- Data models and query languages
- Data quality and data cleansing
- Data science (theory and techniques)
- Context-aware and adaptive information systems
- Constraints extraction from text
- Natural language processing
- Indexing, query processing and optimization
- Information and knowledge extraction
- Information integration
- Information quality
- Graph databases
- Knowledge bases (querying, management, evolution and dynamics)
- Machine learning for knowledge graph construction, completion, refinement
- Machine learning for knowledge and information extraction, for instance, named entity disambiguation, sentiment analysis, relation extraction, or the detection of claims, facts and stances from unstructured documents
- Machine Learning in NLP
- Methodologies, models, algorithms, and architectures for applied data science
- NLP for Digital Humanities
- NLP & Knowledge Graphs
- Privacy, trust and security in databases
- Query processing and optimization
- Question answering over knowledge graphs
- Text databases
Prefered Application Domains (but not limited to).
- Bio-sciences and healthcare
- Urban economy and urban environments
Important Dates. (extended deadline)
- Paper submission :
April 30, 2020. Extended to Sunday, May 3, 2020 (due to requests)
- Notification of acceptance: May 27, 2020
- Camera-ready due: June 5, 2020
- DOING workshop in Lyon: August 25th, 2020 (Invited talk by Marie-Christine Rousset)
- The program will be published in the conference site
DOING workshop intends to accept short (limited to 6 pages) or long (limited to 12 pages) papers. DOING reserves the right to accept only as short papers those papers describing interesting and innovative ideas which still require further technical development. Papers should be written in English, formatted in Latex and present substantially original results. Authors should consult Springer’s authors’ guidelines and use their proceedings templates (you can download the templates available on the bottom of that page).
Accepted papers will be published in the Springer CCIS series and the best papers will be invited to a special issue of the journal Computer Science and Information Systems.
- Cheikh Ba (UGB – Université Gaston Berger, Senegal)
- Javam de Castro Machado (UFC – Universidade Federal do Ceará, Brazil)
- Yi Chen (NJIT – New Jersey Institute of Technology, USA)
- Laurent d’Orazio (IRISA, Université de Rennes, France)
- Vasiliki Foufi (Division of Medical Information Sciences (SIMED), Geneva University Hospitals (HUG), University of Geneva (UNIGE), Switzerland)
- Michel Gagnon (Polytechnique Montréal, Canada)
- Sven Groppe (University of Lubeck, Germany)
- Jixue Liu (University of South Australia, Australia)
- Shuai Ma (Beihang University, China)
- Anne-Lyse Minard-Forst (LLL, Université d’Orléans, France)
- Damien Novel (ERTIM, INALCO, France)
- Fathia Sais (LRI, Université Paris-Sud (Paris-Saclay), France)
- Agata Savary (LIFAT, Université de Tours, France)
- Rebecca Schroeder Freitas (UDESC, Universidade Estadual de Santa Catarina, Brazil)
- Aurora Trinidad Ramirez Pozo (UFPR – Universidade Federal do Paraná, Brazil)
DOING’2020 has counted 17 submissions. Acceptation rate: 50% (8 full papers + 1 short paper)
- Arturo Oncevay, Marco Sobrevilla, Hugo Alatrista-Salas and Andres Melgar. Public Riots in Twitter: Domain-Based Event Filtering during Civil Unrest.
- Ciro M. Medeiros, Umberto S. Costa, Semyon V. Grigorev and Martin A. Musicante. Recursive Expressions for SPARQL Property Paths.
- Jian Lin and Dongming Xie. OMProv: Provenance Mechanism for Objects in Deep Learning.
- Erwan Marchand, Michel Gagnon and Amal Zouaq. Extraction of a Knowledge Graph from French Cultural Heritage Documents.
- Dickson Owuor, Anne Laurent and Joseph Orero. Exploiting IoT data crossings for gradual pattern mining through parallel processing.
- Joshua Amavi, Mirian Halfeld-Ferrari and Nicolas Hiot. Natural Language Querying System through Entity Enrichment.
- Guilherme M. Rocha, Piero L. Capelo and Cristina Dutra De Aguiar Ciferri. Healthcare decision-making over a geographic, socioeconomic, and image data warehouse.
- Dimmy Magalhães and Aurora Pozo. Classification of Relationship in Argumentation using Graph Convolutional Network.
- SHORT PAPER: Damien Alvarez de Toledo, Laurent D’Orazio, Frederic Andres and Maria Leite. Cooking related Carbon Footprint Evaluation and Optimisation.