SHAMAN : A Symbolic and Human-centric view of dAta MANagement
In many application domains (biology, journalism, health, etc.), modern data management applications must allow accessing a large volume of data of varying quality from multiple heterogeneous sources. Furthermore, these applications must allow end-users, who are not IT engineers in general, to specify and efficiently execute complex operations and analyses on data. The goal of Shaman is the design of data management systems that answer these modern application requirements, by relying on techniques from the areas of databases and artificial intelligence, in particular knowledge representation and automated reasoning.
To this aim, the Shaman team investigates three main research axes:
- Efficient data management, both w.r.t. performance and data quality;
- Intelligible data management, in order to help users understanding data, specifying their needs and interpreting/explaining the obtained results;
- Scalable data management, by optimizing the different data management architectures: centralized or massively parallel ones.