French version in the pdf document here.
DIM AI4IDF
Call for doctoral research projects
Co-supervised in Île-de-France
2023
1. Introducing AI4IDF
The Ile-de-France region is home to the world’s largest mathematics community, several of France’s largest computer science laboratories, and a dense industrial fabric in Artificial Intelligence. In this extremely rich context, the four main AI institutes – DATAIA, Hi! PARIS, PRAIRIE and SCAI – have joined forces within the “AI4IDF” program to structure and animate the community, and offer to industrial and international partners a unified vision of the exceptional forces at work. The scientific program of the AI4IDF project aims to deepen knowledge in AI while keeping the human in-the-loop. It is divided into four axes: (1) Learning and optimization, (2) NLP and dialogue with humans, (3) Robotics, movement and interaction with humans, (4) AI in human life: health, education and creation.
2. Submission of files
Applications must be submitted:
- in PDF format only;
- by email to the address ai4idf_call_2023@inria.fr
- before April 14, 2023 at 1 p.m.
Each file must contain:
- the doctoral research project:
- detailing the context, the scientific objective, the justification of the scientific approach
as well as the adequacy with the themes of AI4IDF, - written in English or French,
- not exceeding 3 pages, single-spaced;
- supervision, specifying the role of each supervisor;
- the institutes of the co-supervisors;
- the candidate's CV;
- the doctoral school chosen for the thesis registration if the application is successful.
3. Selection of projects
Eligibility of proposals:
- The subjects proposed must be subject to co-supervision:
✓ - either between two of the four DATAIA institutes, Hi! PARIS, PRAIRIE and SCAI,
✓ - either between one of the four DATAIA institutes, Hi! PARIS, PRAIRIE, SCAI and a laboratory
or Ile-de-France research unit outside these institutes.
- Agreement of the doctoral school mentioned in the file in the event of selection.
The selection committee is made up of 8 scientists with complementary skills in order to provide expertise on the entire scientific program of the project. It is made up half of members from the establishments and organizations of the project, and half of external French or foreign members recognized for their skills in artificial intelligence.
The members of the selection committee will be required to evaluate projects on which they have no conflict of interest.
The evaluation criteria are as follows:
- excellence of the subject and relevance with regard to the scientific program of the DIM AI4IDF;
- excellence of the candidate's profile and suitability of the profile for the project, motivation to join the programme;
- proposed international and/or industrial mobility;
- clarity of the work program envisaged.
Communication of results: Monday, May 15, 2023
4. AI4IDF scientific program
The research program is based on four major areas, each corresponding to major current issues aimed at integrating AI into the human environment as well as possible, in order to improve its existence.
Axis 1 – Learning and optimization: between algorithmic efficiency and theoretical guarantees
Much of the recent advances in artificial intelligence, particularly in computer vision and natural language processing, have been achieved using machine learning algorithms and in particular large-scale optimization. . If the successful implementation of learning in these areas requires “business” expertise (physics of training, image, linguistic) and methods adapted to the corresponding constraints (see axes 2 and 3 below), we can identify scientific questions common to all fields of AI impacted by learning and requiring the combination of new mathematical and computational approaches to be tackled and solved efficiently. These constitute the heart of the “learning and optimization” axis, with three main themes.
- Axis 1.1 Deep learning theory.
- Axis 1.2 Optimal use of resources.
- Axis 1.3 Go beyond pattern recognition: low supervision, structured learning, reinforcement learning.
Axis 2 – NLP and dialogue with humans
The second line of research will focus on areas of AI related to language, and will in particular interact automatic language processing (TAL, or NLP, for "natural language processing") and human-machine dialogue (conversational agents, etc.). It will be organized around four major themes, in strong interaction with each other.
- Axis 2.1 Large-scale language models.
- Axis 2.2 Information extraction and text mining.
- Axis 2.3 Controlled generation.
- Axis 2.4 Transmodal and multimodal tasks.
Axis 3 - Robotics, movement and interaction with humans
For robotics to keep its promises and achieve the expected socio-economic objectives, significant scientific advances therefore remain to be made. They will require both “hardware” experimental platforms benefiting from the latest advances in mechatronics and in the field of sensors and software “intelligence” obtained thanks to fundamental progress in several key areas of AI.
- Axis 3.1 Statistical learning and optimal control.
- Axis 3.2 Perception.
- Axis 3.3 Planning.
- Axis 3.3 Robotics in interaction with human beings.
Axis 4 - AI in human life: examples of health, education and creation
AI has also, through its dissemination as a widely used technique, profoundly modified many other disciplines whose impact in the daily lives of humans is tangible. We will develop these impacts through three areas of application irrigating the various sections of society.
- Axis 4.1. Example of health.
- Axis 4.2. Example of education.
- Axis 4.3. Example of creation.