Sophie Robert



Currently, research activities in the PAMDA's database group (DB-group) are mainly connected to those of DOING.

In 2019, our DB-group created DOING,  an action that assembles researchers interested in transforming data into information and then knowledge. The group starts collaborations with researchers in NTL (natural language processing) and artificial intelligence (AI) to face different challenges:


The idea is to detect the entities and their relationships in textual data and mapping them to a graph database schema. We aim at structuring unstructured textual data to obtain a knowledge graph. The group focuses on the property graph model.

The group has been working for many years on constraint verification and consistent updates (on different contexts such as XML, datalog, RDF, ...). It continues its research on the context of graph databases and proposes to face new challenges such as:  

- Constraint verification and updates on graphs with incomplete information.

Graph databases differ from relational ones by the absence of schema and integrity constraints: data is mixed to metadata. The introduction of constraints imposes understanding that they should be separated into those concerning graph topology and those concerning the values. Different graph models imply different ways of considering these constraints.

- Quality of the database (consistency, reliability, conciseness, privacy protection..).

The group is involved in a collaboration with the SDS team, to propose methods to protect privacy without neglecting database utility. Constraint verifiction has always been an important research topic for our DB-group.

- Intelligent and context-oriented querying of graph databases.

The goal is to propose data science queries, i.e.,  declarative queries that include a request for data analysis on graph databases.