Justine Cassell, PRAIRIE chair, has been elected for a 2 year term. The French National Digital Council is composed of 17 members appointed by the Prime Minister and 4 parliamentarians appointed by the Presidents of the Senate and the National Assembly.

Formed in 2011, the French National Digital Council is an institution which is independent from the Government and whose mission is to formulate opinions on any question relative to the impact of digital technology on society and the economy.

The popular and impressive “automatic language processing” applications can generate so natural texts that they are often indistinguishable from human content. How do they work? What are the risks they may pose? Read the latest article (in French) in “The Conversation” by Thierry Poibeau, PRAIRIE chair.

The students will not only receive a training specific to each PhD project, but they will also take common courses and participate in numerous events. They will all be trained in AI and Machine Learning techniques, in writing articles for science popularization, in Open Science, in cross-disciplinary skills, etc.

The deadline for applications is 26 February. Students admitted will start their PhD work at the beginning of the 2021 academic year.

More information.

AI4theSciences is supported and cofunded by the Horizon 2020-Marie Skłodowska-Curie Actions-COFUND European program.

Every year, the Alexander von Humboldt Foundation grants up to 100 Humboldt Research Awards to internationally leading researchers from abroad, whose fundamental discoveries, new theories or findings  have had a lasting effect on their discipline beyond their immediate research area and who are expected to continue producing outstanding research in the future.

As this award comes with funding to spend some time in research departments in Germany, Jérôme Lang will work at the Universität Düsseldorf and Technische Universität Berlin.

The programs competing for each of these games use the same zero knowledge reinforcement learning algorithms (fundamentally different from the Alpha Zero algorithms), created by Quentin Cohen-Solal and then improved during the PRAIRIE postdoc. This is the first time that the same algorithms have made it possible to win 5 gold medals at the Computer Olympiads.

The presentations from the event are available here.

The AI4Health Winter School gathered more than 450 participants from 30 countries, representing a variety of profiles (60% of students, 20% academics and 20% of private sector professionals).

The General Chair of the event was Emmanuel Bacry, PRAIRIE Chair. Stéphanie Allassonnière, PRAIRIE Chair and Deputy Director, represented PRAIRIE at the opening ceremony.

Dr Cordelia Schmid was awarded the Royal Society Milner Award and Lecture for her work in computer vision and her fundamental contributions to the representation of images and videos for visual recognition. Congratulations!

More information: https://royalsociety.org/science-events-and-lectures/2020/11/milner-lecture/

This €20,000 prize rewards the work of Stanley Durrleman and his team on the development of statistical learning methods and the design of predictive models of the evolution of neurodegenerative diseases such as Alzheimer’s or Parkinson’s. He has developed algorithms capable of predicting the progression of Alzheimer’s disease in patients over the next four years. This tool could play a key role in evaluating the effectiveness of therapeutics for neurodegenerative diseases.

Congratulations!

More information: https://institutducerveau-icm.org/en/actualite/stanley-durrleman-inria-academy-sciences/

6D object pose estimation is one of the key problems in enabling autonomous systems equipped with visual sensors to accurately estimate the positions of objects in the environment in order to grasp and manipulate them. Think, for example, of an assistant robot at home, that can automatically fill-in a dish-washer or a manufacturing robot manipulating objects on the production line.    
The CosyPose approach estimates the 6D pose of multiple known objects in a scene captured by one or more input images with unknown camera viewpoints. The main innovation is a marriage of powerful deep neural networks trained from a combination of synthetic and real images together with multi-view geometric constraints. 

CosyPose: 6D object pose estimation optimizing multi-view COnSistencY. Given a set of RGB images depicting a scene with known objects taken from unknown viewpoints (top), our method accurately reconstructs the scene (bottom) recovering all objects in the scene, their 6D pose and the camera viewpoints. 
The award. The proposed approach achieves state-of-the-art results on multiple benchmarks, doubling the performance of existing methods on the most complex datasets that were beyond the capabilities of previous systems. The method won 5 awards including “The Overall Best Method” in the  6D object pose estimation challenge at ECCV 2020, outperforming other competitors by a significant margin in multiple categories and reaching the performance of methods that use further information from costly depth sensors (which are also sensitive to illumination) while only using non-expensive RGB sensors. The results of the challenge are available as slides and as an overview paper.

Support from the Jean Zay computing cluster. Training the 6D object pose estimation models requires a significant compute on multiple GPUs. Developing the method and winning the challenge would have not been possible without the generous support from the French National Jean Zay computing cluster.

ECCV 2020. The 6D object pose estimation challenge was held in conjunction with the European Conference on Computer Vision, (ECCV) 2020. ECCV is one of the top three computer vision conferences (together with CVPR and ICCV) and is listed among the top 100 most cited journals and conferences over all areas of science by Google Scholar. 

The thesis of Gül Varol covers a wide range of challenging and important problems in human analysis within computer vision. In particular, it defined the state of the art in action recognition from videos with the use of long-term temporal convolutions, and it contributed a synthetic data generation framework for advancing human body shape analysis.
The scientific outcomes of the thesis have been followed up and extended by numerous groups. This has been supported by open sourcing of code and datasets.

Congratulations!

More information: https://ellis.eu/news/ellis-phd-award-2020

Over the last couple of years, automatic translation tools have done significant progress toward improving translation quality. What has driven this improvement in recent years? Listen to the latest podcast with Thierry Poibeau, PRAIRIE chair whose work focuses on Natural Language Processing: https://www.franceinter.fr/emissions/le-code-a-change/la-traduction-automatique-avec-thierry-poibeau

Its goal is to actively participate in the coordination of actions around artificial intelligence in the Paris region, in interaction with other members of the ELLIS network.

Read the press release.

More information on the unit’s website https://ellis-paris.github.io/.

video

L’objectif de ce prix est de dynamiser et de motiver de jeunes chercheurs en les récompensant, et de faire connaître à l’ensemble de la communauté informatique d’excellents travaux de recherche.

Dossier de candidature

Peut candidater tout étudiant ou étudiante ayant soutenu son doctorat d’informatique dans une école ou université française entre le 1er septembre 2019 et le 31 août 2020. Toute candidature devra être explicitement soutenue par le(s) directrice(s) ou directeur(s) de thèse, qui ne peuvent en aucun cas soutenir deux candidats ou candidates.

Date limite de candidature: 14 septembre 2020

Plus d’informations : https://www.societe-informatique-de-france.fr/wp-content/uploads/2020/04/1024-numero-15_Article37.pdf

The mission of the Institut universitaire de France is to promote the development of high-level research at universities and to strengthen interdisciplinarity. Each year it distinguishes a small number of university professors, junior and senior, for their research excellence, as evidenced by their international recognition. The members are appointed for a period of 5 years.

More information (in French): http://www.iufrance.fr

The Postdoctoral Networking Tour in Artificial Intelligence (Postdoc-NeT-AI) is a Project funded by the Federal Ministry of Education and Research (BMBF). Its goal is to connect international early career researchers with the research community in artificial intelligence in Germany.

The Postdoc-NeT-AI offers you the opportunity to participate in one week of on-site visits to leading universities, research institutes and companies; including in depth discussions and networking opportunities.

Here you can find the program of the 11/2020 Tour. Deadline to apply to this year’s Tour is 16 August (midnight CET).

Please find more information on how to apply here.

A key issue addressed by Face Au Virus is how to leverage mobility data in the modeling of the outbreak dynamics.  To complement the health data provided by public authorities, the Initiative gathers additional data, provided by private companies (including Facebook and Roofstreet), which enable the monitoring of the evolution of mobility according to its various characteristics.

Face Au Virus pays great attention to ensure that the use of data is carried out with respect for fundamental rights and freedoms and in no way aims to identify individual behavior.  The initiative does not collect any personal data and only processes statistical summaries of mobility data.

More information (in French): https://www.psl.eu/initiative-scientifique-covid-19-face-au-virus

The Longuet-Higgins Prize is presented annually by the IEEE Pattern Analysis and Machine Intelligence (PAMI) Technical Committee at each year’s Conference on Computer Vision and Pattern Recognition for fundamental contributions in computer vision. The award recognizes CVPR papers from ten years ago with significant impact on computer vision research.

The prize is named after theoretical chemist and cognitive scientist H. Christopher Longuet-Higgins.

More information (in French): https://www.inria.fr/fr/jean-ponce-recoit-le-prix-longuet-higgins

The fellowships are available for M1 or M2 years and are awarded on academic criteria. Please find more information here.

The web platform challengedata.ens.fr includes the presentation of problems, and ensures exchanges of data and automatic evaluation of the results of participants. Lecturers can set up an account to select challenges adapted to their course and get access to a dashboard which monitors student activity during the course. The students can upload a report, which is downloadable by the pedagogical supervisor.

Have a look at the videos presenting the challenges: http://www.college-de-france.fr/site/stephane-mallat/Challenges-2020.htm

More information about Challenge Data (in French) here

In order to participate take a look at general documentation and FAQ.

Marc Lelarge is the co-founder at Pluton AI and Deep Learning lecturer at Ecole Polytechnique and Ecole Normale Supérieure.

The course is designed with a wide target audience in mind: software developers who want to delve deeper in AI, scientists who want to want to acquire practical experience, or entrepreneurs who seek state-of-the-art technology to power their next idea.

More information: https://pluton.ai/course/