While AI is becoming increasingly ubiquitous and entering areas where it was still unknown until recently (human resource management, administrations, fintech, fundamental sciences, e-health, justice, industry 4.0, etc.), it becomes clear that a relationship of trust must be established between users and the latter. Whether they are experts or not, people facing AI are entitled to expect certain guarantees (reliability, respect for data confidentiality, stability and consistency of decisions, etc.).
The work package 2 “Confidence Mechanisms for AI” seeks to address this need from different angles: federated learning based on blockchain technologies such as Trusted Coalitions for Blockchain Distributed Learning (TCLearn) and (BFAs) ; inductive logic for predictive justice; guarantees such as stability, the choice of good metaphors (especially in deep learning), self-assessment and certification; robustness to variations in objectives, to the quality of supervision and to data not-basedviews, including internal representations; the interpretation of models via mechanisms of hybridization, distillation, constraints and compromise complexity-optimality; interaction for example for infovis and multi-agent systems for by-effects guaranteesdesign for robots and swarms. All these different angles will therefore be studied under this line of research.
The work package is coordinated by Prof. Benoît Frénay and Dr. Rebecca Marion at UNamur. A call was made for a first WP2 launch meeting in March, after the start of the second semester. The aim will be to identify areas of interest and clarify the tasks that ARIAC researchers are working on. Regular meetings will then be scheduled (e.g. every two months) to ensure cohesion and visibility on the progress of the different teams and also to have a view on how the research work and sub-tasks of major challenges can be articulated. A first workshop will be organized as part of the 33rd International Francophone Conference on Human-Machine Interaction (HMI'22) to be held at the University of Namur from 5 to 8 April 2022, Namur (https://ihm2022.afihm.org/en/ ) with the theme “Human-Machine Interaction and Explicability in Artificial Intelligence”. WP2 speakers (and more broadly ARIAC) will be invited to contribute.
A total of 20 researchers, 14 of whom are funded by the ARIAC project, are currently working on this area of research. We are pleased to have been able to attract expertise from several partners. Most ARIAC researchers are affiliated with UNamur, UCLouvain and CETIC.