ISAE-SUPAERO stakes out its position in the race toward artificial intelligence
Available online :
By welcoming the Institut d’Automne en Intelligence Artificielle (IA² - Autumn Institute of Artificial Intelligence) to its campus in October 2018, ISAE-SUPAERO has been pursuing its involvement in the development of the different Artificial Intelligence (AI) disciplines.
ISAE-SUPAERO is also participating in the ANITI (Artificial and Natural Intelligence Toulouse Institute), project promoted by the University of Toulouse, which was preselected by an international panel last November for its project for an interdisciplinary institute dedicated to artificial intelligence (3IA).
Artificial intelligence is gaining momentum in the fields of transportation, environment and healthcare. It is raising the stakes for all socioeconomic, territorial and academic partners in the region targeted by this project for an interdisciplinary institute.
The Institut d’Automne en Intelligence Artificielle (IA² - Autumn Institute of Artificial Intelligence)
IA² was organized jointly by ISAE-SUPAERO and the CNRS. This year’s topic dealt with artificial intelligence and robotics, a research field in line with the work carried out on drones and robots at the Institute. AI research activities at ISAE-SUPAERO are organized by the Complex Systems Engineering Department (DISC) and focus on automatic learning. DISC also uses artificial intelligence techniques for proof of programs and embedded software certification.
The courses offered to the approximately sixty participants covered a wide variety of topics such as deep-learning for vision, planning with uncertainty, multi-robot coordination, AI ethics, robotics, autonomous vehicles, etc.
We asked Sébastien Konieczny, Director of Research in Artificial Intelligence at the Lens Computer Science Research Lab (CRIL - Université d’Artois) and scientific director at the Institute (IA²), to tell us more about the institute’s objectives.
"The institute’s goal is to present the main techniques used in AI. AI is not a homogenous whole. Our goal in AI research is not to create robots that can think and move, but rather to model intelligent behaviors. These behaviors vary widely in nature and call upon different techniques.
We look into techniques dealing with perception, vision and speech recognition. There is also learning, learning new things, managing to use what is learned to come to new conclusions, and therefore reasoning.
There are also decision-making issues. What are the right decisions to be made once all the data are available?
Lastly, we look into questions of action, i.e. what to do once the right decisions have been made and the right reasoning has been applied – what action should be taken upon the world?
Artificial intelligence is a wide set of different techniques and research has to be carried out on each of these points.
AI research is a group of several topics that must be combined to develop robots that will really be useful one day".
.
A research strategy in artificial intelligence developed nationwide
In 2017, Sébastien Konieczny also participated in the parliamentary mission on artificial intelligence ordered by the French Prime Minister, Édouard Philippe, and headed by Cédric Villany. The Villany report was the product of the call for expressions of interest that led to the preselection of four dossiers out of the twelve presented: Grenoble (MIAI@Grenoble_Alpes), Paris (Prairie), Nice-Sophia Antipolis (3IA Cote d’Azur) and Toulouse (ANITI).
ANITI’s goal is to:
- develop a new generation of hybrid artificial intelligence that will provide guarantees in terms of reliability and the capacity to explain and interpret the results of algorithms while ensuring social acceptability and economic viability;
- double the number of artificial intelligence graduates by the year 2023;
- create some forty chairs and a graduate school.
- The procedure will then lead to ruling on the definitive labelling of those sites that will welcome a 3IA at the start of 2019. Each institute will then be allocated a share of the budget allowance granted by the State through the ANR (National Research Agency).