Cognitive AI for decision support and multi-objective optimization

We are seeking a consortium in need of a technology partner capable of producing and integrating intelligible predictive models (more effective than only explainable) to facilitate decision-making and/or the search for optimal solutions. Model intelligibility is essential for prescriptive systems in medicine.

We use augmented cognitive AI (XTRACTIS) that produces intelligible (transparent) predictive models through the discovery of logical rules (knowledge).

XTRACTIS induces rules from structured data (biomarkers, epigenomes, ...). For image and temporal signal modalities, it is necessary to first extract the features relevant to the domain expert (preprocessing).

XTRACTIS can meet the requirements of the EU AI ​​Act for high-risk systems.

Another fundamental characteristic of this approach is the ability to invert models to search for solutions (drug formulation, prescription) through multi-objective optimization.