ANR PRCE FIDD
Facilitated Exploration: Interactive Constraint-Driven Data Mining
The FIDD project disrupts data exploration through a continuous learning loop. It acquires user-specific constraints using Constraint Acquisition. Efficient, declarative model revision for pattern extraction, employing symbolic AI techniques (e.g., CP, SAT), ensures optimal problem resolution. The project goal is to create an iterative process in which users interact dynamically, enabling active learning and adaptation. Advanced machine learning understands evolving user needs, while constraint reasoning continually refines the pattern extraction model. Aligned with user-centric interactive data mining, FIDD project captures user feedback declaratively, optimizing efficiency, and facilitating dynamic model revisions. This novel fusion of Constraint-Driven Data Mining and Constraint Acquisition represents a pioneering effort in the field.