How can precise, relevant and useful knowledge be produced in a highly targeted industrial or commercial context from massive volumes of heterogeneous data from multiple and varied sources? This is the challenge that SystemX and its partners intend to overcome by combining their multi-sector expertise and hybridizing recent approaches to digital learning and symbolic AI techniques with business knowledge.
SystemX is launching the “Business Semantics for Multi-Source Data Mining” (SMD) project, the third project in its “Artificial Intelligence and Augmented Engineering” (IA2) research program. This 48-month collaborative R&D project brings together five companies (Airbus Defence and Space SLC, Apsys, Ecosys, EDF, RTE) and one research institution (CentraleSupélec) to develop tools that hybridize symbolic AI and learning AI for the intelligence and knowledge management sectors, to build and exploit knowledge on heterogeneous multi-source data, and to support decision making in static or dynamic environments.
The project aims mainly to design a system which hybridizes symbolic AI and recent learning methods (Deep Learning) with business knowledge, for constructing new and intelligible knowledge based on volumes of unstructured, heterogeneous, and multi-source data, and for making relevant recommendations for better decision making.
“The project came about due to the real and recurrent observation that any industrial company could face during its digital evolution, in the advent of a significant volume of highly heterogeneous, unstructured and multi-source data. With the increase in computing power of computers, we are increasingly able to retrieve information. However, these attributes alone do not provide sufficient understanding comparable to that of humans, hence the importance of coupling several approaches. The SMD project aims to remove a major obstacle concerning the hybridization of knowledge representation and reasoning approaches (symbolic AI) with recent approaches in Artificial Intelligence (for example, le Deep Learning) for the analysis of heterogeneous data”, explains Sana Tmar, project manager of SMD, IRT SystemX.
