eXplica-IA: Designing eXplainable AI Systems which interact with people in Natural Language

In the last years, research in Artificial Intelligence reached high levels of excellence in science and technology. As a consequence, milestones which would be considered science fiction some years ago are becoming present realities. The use of the term “smart” is becoming more and more pervasive, as we now have smart cities, smart factories, smart houses, smart cars, smart mobiles, etc. In our daily life, we are surrounded by devices that include some level of intelligence without being conscious of it. Artificial Intelligence is boosting a social revolution with the potential of changing the way we live. For this reason, the number of studies dealing with how to build an Artificial Intelligence more human and accessible to the citizens is continuously increasing. As the interaction between people and devices increases, technological aspects mingle with ethical and legal questions which have to be addressed.

Objectives

The main objective of the eXplica-IA project is to answer the scientific challenge of designing and developing a new generation of algorithms and Artificial Intelligence techniques which can provide self-explanatory capabilities to Intelligent Systems; that is, with the capacities to explain their behavior and decisions in Natural Language clearly and without linguistic ambiguities, both to expert and non-expert users.

New algorithms and techniques will be validated against functional and non-functional requirements, in real use cases of interaction between people and machines. With the aim of maximizing the diffusion to the scientific community and to the industry, we will develop an open source platform integrating the technologies developed during the project. Moreover, this software will guide the user in the intuitive design of explainable intelligent systems and their rigorous evaluation.