With PAISS, artificial intelligence "goes back to school" in Grenoble
From 2 to 6 July 2018, the Inria Grenoble centre will welcome researchers, students and members of industry to its PAISS summer school, dedicated to artificial intelligence - an initiative, led by Inria and Naver Labs, in order to further knowledge and stimulate interactions.
The AI for Humanity summit, organised in the spring by the French President of the Republic, the report by the French MP Cédric Villani, but also recurring articles or reports on the theme...which topic other than artificial intelligence intrigues not only the academic, industrial and political spheres but also the general public both in France and internationally? For the THOTH team from Inria Grenoble - which specialises in visual recognition and machine learning - this context is particularly conducive to debate and questioning. And so the idea was born of breathing new life into the series of summer schools that ended in 2013, and which was organised every two years alternately in the Inria centres of Paris and Grenoble. « Artificial intelligence is certainly omnipresent, but do we really know where the state-of-the-art is at? What are its perspectives in concrete terms? We felt it was important to be able to bring together scientific and industry players to take stock of the subject. That is why we decided to launch PAISS - PRAIRIE Artificial Intelligence Summer School - in partnership with the laboratory Naver Labs, set up near Grenoble. The principle of alternating between Paris and Grenoble is, moreover, retained, however this time it takes place every year in order to follow developments in the field as closely as possible » , Julien Mairal, researcher with the THOTH team, explains.
One watchword: diversity
And so the teams have managed the amazing feat of organising this event in just six months, which will take place over five days on the Inria premises in Montbonnot, on the outskirts of Grenoble. Julien Mairal explains: « Artificial intelligence is, by nature, multidisciplinary. Our aim was therefore to show to what extent. This is why the speeches will focus on different sub-fields such as visual recognition, machine learning, mathematical optimisation, robotics and even the cognitive sciences. » This same priority of diversity influenced the selection of the participants - initially 700 applying for only 200 places. In order to provide a cross-section of approaches and viewpoints, three profiles are represented: researchers, industry players and students. Moreover, Naver Labs has put in place a system of grants for students to enable those who do not necessarily have the means to attend the event.
An exceptional cast
On the programme: 15 lectures, each lasting 1½ hours, given by internationally renowned experts, representatives of both the academic world and industry. They notably include Yann Le Cun and Léon Bottou from Facebook, Rémi Munos from the Deepmind laboratory, but also Cordelia Schmid, Jean Ponce, Martial Hébert from Carnegie Mellon University, Andrew Zisserman from the University of Oxford... « The presentations will be very high-level thanks to the great expertise of the speakers », Julien Mairal summarises. Three practical sessions will also be organised by Naver Labs, Criteo and Facebook to supplement the approaches.
The quality of the scientific programme is certainly one of the keys to the success of this event. However, it is not the only one. Indeed, success will also depend on the discussions - synergies, even - that could develop between the participants. The programme therefore includes a welcome reception on the first day at La Bastille in Grenoble, whilst during the week Naver Labs will be organising a gala evening in their unusual premises which include a castle and large gardens. In parallel, a "posters" session will also enable the students to present their work and meet industry players. « Over time, these interactions can lead to new collaborations. Their vocation is also to showcase our scientific environment and our regional and national appeal with regard to artificial intelligence », Julien Mairal concludes.