ANITI, Artificial and Natural Intelligence Toulouse Institute

The interdisciplinary institute in artificial intelligence of Toulouse, named the Artificial and Natural Intelligence Toulouse Institute (ANITI), has been selected to be one of four institutes spearheading research on AI in France. The challenge is to make Toulouse one of the world leaders in artificial intelligence in research, education, innovation and economic development. The strategic application sectors targeted by the project are mobility and transportation, and robotics/cobotics for the industry of the future.

The ambition of the ANITI project is to develop a new generation of artificial intelligence called hybrid AI, combining data-driven machine learning techniques with symbolic and formal methods for expressing properties and constraints and carrying out logical reasoning.

This approach will provide better guarantees in terms of reliability, robustness and the ability to explain and interpret the results of the algorithms used, while ensuring social acceptability and economic viability. Such guarantees are required by many applications targeted by the project, such as autonomous vehicles of the future. Starting operations this autumn, ANITI will bring together more than 200 researchers from universities, engineering schools, scientific and technological research organizations, and about thirty companies in the Toulouse region.

investissement d'avenir.pngANITI is coordinated by the University of Toulouse: Université fédérale Toulouse Midi-Pyrénées within the framework of France1s << lnvesting for the Future – PIA3 >> program, with the support of the Occitanie Region, the Toulouse Metropole, and the SATT Toulouse Tech Transfer.

+ 50 PARTNERS

ogos partenaires ANITI - 3IA

Actia Automotive, Aerospace Valley, AIRBUS, Altran, Atos Integration, Brainkey, Caisse d’Épargne, Capgemini, CERFACS, CGI, Centre hospitalier universitaire de Toulouse, CLS, CNES, CNRS, Continental, CS Systèmes d’information, EDF, ENAC, Groupe BRL, Groupe Renault, IBM, ICAM, IMT Mines d’Albi, Inra, INSA Toulouse, Inserm, INU Champollion, IRD, IRT Saint-Exupéry, ISAE-SUPAERO, IVADO, Latécoère, Liebherr, Linagora, Météo-France, NXP, ONERA, Pierre-Fabre, Quantmetry, Qwant, Scalian, Siemens, Sopra Steria, Syngenta, TBS, THALES, Toulouse INP, Toulouse INP-ENIT, Université fédérale Toulouse Midi-Pyrénées, Université Toulouse Capitole, Université Toulouse – Jean Jaurès, Université Toulouse III – Paul Sabatier

ANITI
2 strategic application sectors targeted : mobility and transportation & robotics/cobotics for the industry of the future
200+ researchers
3 intégrative programs
22 research chairs
50+ partners including some thirty companies

 

.

TARGET BUDGET OVER FOUR YEARS
100 M€ academia, industry, PIA3 investment programme, institutions
including
24 M€  Occitanie Région 
4 M€ Toulouse Métropole

 

.

UNIVERSITÉ FÉDÉRALE TOULOUSE MIDI-PYRÉNÉES
100,000+ students
31 universities, schools & research entities
1,000+ training courses: BA/M/PhDs
145 research laboratories and entities
5th largest concentration in France of ERC researchers

TOULOUSE & ITS REGION
5th largest concentration of researchers in France
6800 public sector researchers
2nd  largest creator of startups in France (INSEE 2018)

ANITI, Artificial and Natural Intelligence Toulouse Institute

The interdisciplinary institute in artificial intelligence of Toulouse, named the Artificial and Natural Intelligence Toulouse Institute (ANITI), has been selected to be one of four institutes spearheading research on AI in France. The challenge is to make Toulouse one of the world leaders in artificial intelligence in research, education, innovation and economic development. The strategic application sectors targeted by the project are mobility and transportation, and robotics/cobotics for the industry of the future.

The ambition of the ANITI project is to develop a new generation of artificial intelligence called hybrid AI, combining data-driven machine learning techniques with symbolic and formal methods for expressing properties and constraints and carrying out logical reasoning.

This approach will provide better guarantees in terms of reliability, robustness and the ability to explain and interpret the results of the algorithms used, while ensuring social acceptability and economic viability. Such guarantees are required by many applications targeted by the project, such as autonomous vehicles of the future. Starting operations this autumn, ANITI will bring together more than 200 researchers from universities, engineering schools, scientific and technological research organizations, and about thirty companies in the Toulouse region.

ANITI is coordinated by the University of Toulouse: Université fédérale Toulouse Midi-Pyrénées within the framework of France1s << lnvesting for the Future – PIA3 >> program, with the support of the Occitanie Region, the Toulouse Metropole, and the SATT Toulouse Tech Transfer.

+ 50 PARTENERS

ogos partenaires ANITI - 3IA

Actia Automotive, Aerospace Valley, AIRBUS, Altran, Atos Integration, Brainkey, Caisse d’Épargne, Capgemini, CERFACS, CGI, Centre hospitalier universitaire de Toulouse, CLS, CNES, CNRS, Continental, CS Systèmes d’information, EDF, ENAC, Groupe BRL, Groupe Renault, IBM, ICAM, IMT Mines d’Albi, Inra, INSA Toulouse, Inserm, INU Champollion, IRD, IRT Saint-Exupéry, ISAE-SUPAERO, IVADO, Latécoère, Liebherr, Linagora, Météo-France, NXP, ONERA, Pierre-Fabre, Quantmetry, Qwant, Scalian, Siemens, Sopra Steria, Syngenta, TBS, THALES, Toulouse INP, Toulouse INP-ENIT, Université fédérale Toulouse Midi-Pyrénées, Université Toulouse Capitole, Université Toulouse – Jean Jaurès, Université Toulouse III – Paul Sabatier

  Projects

SCIENTIFIC PROJECT

The scientific project is structured around three integrative programs (IPs), which will develop innovative solutions to address challenges raised by our application domains using theoretical advances in core AI scientific areas.

Acceptability, Fair representative data for AI
This IP addresses various facets of the acceptability of systems integrating AI algorithms from social, economical, legal or ethical points of view. This includes issues about data that can affect AI algorithms. We will propose new ways of handling data to address data bottlenecks and data biases that can hamper AI systems.

Certifiable AI toward autonomous critical Systems
This IP will develop new methods, models and tools based on hybrid AI, to support the design and validation of critical autonomous systems for which strong guarantees are required, (e.g., by certification authorities in aeronautics). This program will strengthen and implement the momentum initiated by the IRT-Saint Exupéry on this topic.

Assistants for design, decision, and optimized Industry processes
This IP will develop new AI methods to aid human decisions. This program will design advanced AI assistants to increase the performance of design, decision and industrial production related activities. This will lead to i) the design of cognitive assistants with advanced dialogue and interaction skills, ii) the monitoring of complex systems in order to model their behaviour, predict their evolution, and anticipate corrective actions, and iii) the design of autonomous mobile robots with the ability to interact with humans, cognitively and physically, to perform complex tasks in a collaborative manner.

In total, the project aims to fund more than thirty research chairs, of which about ten would be supported by researchers from international laboratories and universities (e.g. MIT or Brown University in the United States). The project will also promote international mobility and collaboration to attract outstanding students and the best experts to address the challenges tackled by the project.

EDUCATION AND TRAINING PROJECT

The ambition is to become a world leader in hybrid AI education and to double the number of students trained in AI by 2023.

Aniti will start at high school to increase AI awareness among high school students, especially girls. The efforts at the high school level will increase flow of students in undergraduate programs providing mathematical and computer skills at the BA level, one of ANITI’s priorities. AI modules throughout BA programs will be added and new advanced programs will be created. At the Master's and PhD level, Aniti will create a new AI Graduate school, focusing on hybrid AI, with worldwide visibility to attract talented students.

The project will also address the lack and urgent need in industry for AI qualified personnel, by developing apprenticeship programs and devoting significant effort to continuing education.

A single portal entry for continuing education for the Toulouse site will be offered, with programs tailored to different levels from managerial awareness to data scientist familiarity even expertise in hybrid AI.

The solutions will be adaptable to SMEs as well as big industrial groups.

HIGH-SCHOOL
. New doctoral program on hybrid AI
. Research oriented program 1st year on (acd/indus.projects)
. Interdisciplinarity (cognitive science, Ethics, Law, Economics, humanities)

BA
. Increase the flow of students in existing AI programs
. Create new BA programs - "advanced AI" (Maths, CS)
. Include AI modules throughout BA programs

GRADUATE SCHOOL MA/PHD
. Raise AI awareness especially among HS girls

le projet culture scientifique

SCIENTIFIC CULTURE

Several actions to disseminate AI scientific culture will be planned, drawing on local strengths.

ECONOMIC DEVELOPMENT PROJECT

ANITI will set up interfaces with key players in the innovation and technology transfer ecosystem to promote the results from research projects and study the opportunities for exploiting and using these results. The aim is to foster the creation of disruptive technologies paving the way for new economic prospects for partners.

The project intends to drive the creation of around one hundred startups for student entrepreneurs, industrial players and academics between now and 2023. Lastly, ANITI will also allocate resources to the transfer of integrative programme results to industrial partners, particularly SMEs, with the support of our partners clusters (Aerospace Valley, etc.).

> Start-up creation

• Via ANITl's Innovation andBusiness Committee
• Up to 1 M€/year
dedicated byToulouse Tech Transfer(early stage funding)
• Pre-incubation and incubationmanagement
in liaison with innovationclusters (e.r., Aerospace Valley)
public and private incubators .

Rapid dissemination of new technological possibilities
to ANITI partners via IPs

SCIENTIFIC PROJECT

The scientific project is structured around three integrative programs (IPs), which will develop innovative solutions to address challenges raised by our application domains using theoretical advances in core AI scientific areas.

Acceptability, Fair representative data for AI
This IP addresses various facets of the acceptability of systems integrating AI algorithms from social, economical, legal or ethical points of view. This includes issues about data that can affect AI algorithms. We will propose new ways of handling data to address data bottlenecks and data biases that can hamper AI systems.

Certifiable AI toward autonomous critical Systems
This IP will develop new methods, models and tools based on hybrid AI, to support the design and validation of critical autonomous systems for which strong guarantees are required, (e.g., by certification authorities in aeronautics). This program will strengthen and implement the momentum initiated by the IRT-Saint Exupéry on this topic.

Assistants for design, decision, and optimized Industry processes
This IP will develop new AI methods to aid human decisions. This program will design advanced AI assistants to increase the performance of design, decision and industrial production related activities. This will lead to i) the design of cognitive assistants with advanced dialogue and interaction skills, ii) the monitoring of complex systems in order to model their behaviour, predict their evolution, and anticipate corrective actions, and iii) the design of autonomous mobile robots with the ability to interact with humans, cognitively and physically, to perform complex tasks in a collaborative manner.

In total, the project aims to fund more than thirty research chairs, of which about ten would be supported by researchers from international laboratories and universities (e.g. MIT or Brown University in the United States). The project will also promote international mobility and collaboration to attract outstanding students and the best experts to address the challenges tackled by the project.

EDUCATION AND TRAINING PROJECT

The ambition is to become a world leader in hybrid AI education and to double the number of students trained in AI by 2023.

Aniti will start at high school to increase AI awareness among high school students, especially girls. The efforts at the high school level will increase flow of students in undergraduate programs providing mathematical and computer skills at the BA level, one of ANITI’s priorities. AI modules throughout BA programs will be added and new advanced programs will be created. At the Master's and PhD level, Aniti will create a new AI Graduate school, focusing on hybrid AI, with worldwide visibility to attract talented students.

The project will also address the lack and urgent need in industry for AI qualified personnel, by developing apprenticeship programs and devoting significant effort to continuing education.

A single portal entry for continuing education for the Toulouse site will be offered, with programs tailored to different levels from managerial awareness to data scientist familiarity even expertise in hybrid AI.

The solutions will be adaptable to SMEs as well as big industrial groups. Finally, several actions to disseminate AI scientific culture will be planned, drawing on local strengths.

ECONOMIC DEVELOPMENT PROJECT

ANITI will set up interfaces with key players in the innovation and technology transfer ecosystem to promote the results from research projects and study the opportunities for exploiting and using these results. The aim is to foster the creation of disruptive technologies paving the way for new economic prospects for partners.

The project intends to drive the creation of around one hundred startups for student entrepreneurs, industrial players and academics between now and 2023. Lastly, ANITI will also allocate resources to the transfer of integrative programme results to industrial partners, particularly SMEs, with the support of our partners clusters (Aerospace Valley, etc.).

le projet culture scientifique

SCIENTIFIC CULTURE

Several actions to disseminate AI scientific culture will be planned, drawing on local strengths.

  Chairs

IPA - INTEGRATIVE PROGRAM A: ACCEPTABILITY , FAIR REPRESENTATIVE DATA FOR AI

This IP addresses various facets of the acceptability of systems integrating Al algorithms fromsocial, economical, legal or ethical points of view. This includes issues about data that can affectAl algorithms. We will propose new ways of handling data to address data bottlenecks and databiases that can hamper Al systems.

Moral AI

SHS-Socio-Psycho: Jean François Bonnefon (DR CNRS)
This project will examine and quantify judgments of ordinary citizens concerning solutions to problems that involve tradeoffs. This will bring quantitative methods to bear on well known ethical dilemmas like the Trolley problem, which are very relevant for the design of autonomous vehicles.
Email: jfbonnefon@gmail.com
https://sites.google.com/site/jfbonnefon/home

Law, Accountability and Social Trust in AI

Legal Issues: Céline Castets-Renard (PR UT1C)
The goal of this chair is to investigate how a legal framework for making AI programs properly accountable. There are important legal issues–consumer protection, liability, and insurance—that need work before AI can gain full social trust and that is crucial to making AI widespread. French law also has a requirement on the explainability of an administrative decision, but this notion will need clarification and refinement in the face of the use of AI methods in administration
Email: castetsrenard@gmail.com
https://www.castetsrenard.org

The effects of AI on competition in the marketplace

AI and market competition: Bruno Jullien (DR CNRS)
This research program aims at fostering our theoretical and empirical understanding of the economics of information services using AI, with a special emphasis on the impact of AI on competition. Do AI based pricing algorithms inevitably or tend to lead to collusion, price-fixing or predatory pricing? If so, how can we control for that. Do AI programs lead to de facto monopoly behavior amongst would be competitors? How also do data driven mergers affect competitivity? Answering such questions is important for the eventual acceptance by the public of ubiquitous AI
Email: br.jullien@gmail.com
https://www.tse-fr.eu/fr/people/bruno-jullien

Fair & Robust Methods in Machine Learning

Fairness/Robustness PI: Jean-Michel Loubes (PR UT3, IMT) – also in IP-B
This project will analyse fair learning and bias using tools from statistics and optimal transport theory and contribute to explaining ML program behavior, anomaly detection and making ML methods more robust. It will provide means for removing biases from both data and machine learning algorithms by adding certain constraints to the learning process, for instance. By isolating the effects of a known bias and observing changes in a program’s behavior after a certain bias has been removed, this work will also contribute to explaining ML program behavior and may figure also in anomaly detection and in determining whether a ML method is robust or not.
Email: loubes@math.univ-toulouse.fr
https://perso.math.univ-toulouse.fr/loubes/

Empowering Data-driven AI by Argumentation and Persuasion

Hybrid Argument: Leila Amgoud (DR CNRS, IRIT)

The project will examine argumentation structures extracted from ML program behavior to support the conclusions they arrive at, thus enhancing the explainability of data-driven AI. Another more general thread of this project will be to uncover the links of persuasion, biases, learnability and argumentation, linking with the projects of Loubes and Castet-Renard.

Email: amgoud@irit.fr

https://www.irit.fr/~Leila.Amgoud

Developing AI to Improve Global Governance

Augmented Society: Cesar Hidalgo (External, MIT, USA)
The goal of this research program is to advance the development of big data and AI tools to serve the general public and promote data driven decision making—in particular tools like public data distribution platforms, public data that have added content from computer vision and natural language processing, digital twins for daily decision making, and AI ethics.
Email: hidalgo@mit.edu
http://chidalgo.org/

Data-driven approximate Bayesian computation for fusion-based inference from heterogeneous (remote sensing) data

Remote Sensing Data Analysis PI: Nicolas Dobigeon (PR, INP)
This project will apply approximate Bayesian computation (ABC) to problems with algorithms for extracting hidden properties (described in a latent space) in multi source, multi-scale and multi temporal data that are often heterogeneous and thus have no straightforward physical model that offers a general descriptive framework. ABC methods offer an approximate descriptive framework for various generative models including various deep learning methods (auto encoders, generalized adversarial networks).
Email: nicolas.dobigeon@enseeiht.fr
http://dobigeon.perso.enseeiht.fr

AI for physical models with geometric tools

Simulating Complex Environments PI: Fabrice Gamboa (PR UT3)
This project will look at complex computer simulations, which are used to model complicated physical, chemical or biological phenomena, and seek to improve their analysis by using the geometry or topology of the parameter space (of the computational model) or the data, with application to various data driven deep learning models.
Email: fabrice.gamboa@math.univ-toulouse.fr
https://www.math.univ-toulouse.fr/~gamboa/

IP B - INTEGRATIVE PROGRAM B: CERTIFIABLE Al TOWARD AUTONOMOUS CRITICAL SYSTEMS

This IP will develop new methods, models and tools based on hybrid Al, to support the design and validation of critical autonomous systems for which strong guarantees are required, (e.g., by cer tification authorities in aeronautics). This program will strengthen and implement the mo­mentum initiated by the IRT-Saint Exupéry on this topic

Efficient algorithms and Data Assimilation for Computationally Efficient Constrained Advanced Learning

Data Assimilation and Machine Learning: Serge Gratton (PR INP)
This project will design gradient based embeddable algorithms, that are provably convergent to 2nd order stationary points, with a provable low complexity. To reach this aim we will rely on ideas that have proved efficient in more general optimization settings for problems with data: domain decomposition, multilevel methods, inexact computing, nonlinear preconditioning, randomized and quasi-static methods to escape saddle points. The projectwill also explore data assimilation approaches, which provide a Bayesian framework for learning under physical constraints along a time dimension.
Email: serge.gratton@enseeiht.fr
http://gratton.perso.enseeiht.fr

Deep Learner Explanation & VERification

Hybrid, Subsymbolic > Symbolic: Joao Marques-Silva (external, University of Lisbon, PT)
This project envisions two main lines of research, concretely explanation and verification of deep ML models. It will build on the remarkable progress made by automated reasoners based on SAT, SMT, CP, ILP solvers (among others) to further explainable and robust data driven AI (Hybrid AI for proving robustness for neural networks). The reasons for these successes include improved solver technology, more sophisticated encodings, and also by exploiting key concepts that include abstraction refinement, symmetry identification and breaking among others.
Email: jpms@ciencias.ulisboa.pt
http://www.di.fc.ul.pt/~jpms/

Optimization for ML and the Christoffel function for data analysis
Many ML applications such as unsupervised clustering or deep learning, are formulated as non-convex problems. In addition, from a complexity point of view, we often face average case problems where data are drawn from distributions, and a better understanding of such situations is required. This project will be looking at polynomial optimisation using approximation methods for nonconvex search spaces and various functions for data analysis.
Email: lasserre@laas.fr
http://homepages.laas.fr/lasserre/
Pushing the frontier of nonconvex optimization to more general settings and understanding why it works
The main objectives are to extend the scope of algorithms that can cope with nonconvexity and the curse of dimensionality by exploiting data information and structures, to analyze their mathematical properties, to identify the determining factors of their numerical complexity, to improve their performance, and to apply them to solve high impact applied problems. Many ML problems require solving nonconvex optimization, e.g., clustering problems; dimension reduction paradigms such as sparse PCA (Principal Component Analysis), Nonnegative Matrix Factorization, to name just a few. In all these instances, the problems are highly nonconvex, huge scale and even nonsmooth, and are the source of challenging open questions.
Email: teboulle@post.tau.ac.il
http://www.math.tau.ac.il/~teboulle
Game Theory, Convergence for Generalized Adversarial Nets and other ML architectures

Games and Adversarial Nets PI: Jérôme Renault (PR UT1C)
This project will formally study and prove properties about of one the most complex learning architectures: Generalized Adversarial Networks (GANs), and also about complex interactions of autonomous AI systems, using stochastic games.
Email: jerome.renault@tse-fr.eu
https://sites.google.com/site/jrenaultsite

Large scale optimization for AI

Large scale optimization for AI PI: Jérôme Bolte (PR UT1C)
This chair will study convergence properties/rates, global optimization and error bounds, design/optimization of underlying geometrical structure, optimization of adversarial models. In addition, it will explore the modeling and algorithms for large and autonomous systems using bio-inspired models. Particle swarm algorithms, which will arise out of bio inspired models for group decision/ action beyond the capacities of extant models like boids, will be considered and optimisation techniques for them
Email: jbolte@ut-capitole.fr
https://www.tse-fr.eu/fr/people/jerome-bolte

Fair & Robust Methods in Machine Learning

Fairness/Robustness PI: Jean-Michel Loubes (PR UT3, IMT) – also in IP-A
This project will analyse fair learning and bias using tools from statistics and optimal transport theory and contribute to explaining ML program behavior, anomaly detection and making ML methods more robust. It will provide means for removing biases from both data and machine learning algorithms by adding certain constraints to the learning process, for instance. By isolating the effects of a known bias and observing changes in a program’s behavior after a certain bias has been removed, this work will also contribute to explaining ML program behavior and may figure also in anomaly detection and in determining whether a ML method is robust or not.
Email: loubes@math.univ-toulouse.fr
https://perso.math.univ-toulouse.fr/loubes/

New certification approaches of AI based systems for civil aeronautics

Certification PI: Claire Pagetti (ONERA)
The work proposed consists in identifying lack of certification standards for new phenomena and proposing new approaches to help develop and certify AI applications, e.g., integrating the notion of algorithm failure, non deterministic and unpredictable behaviour, run time services to detect faulty behavior in sophisticated AI systems.
Email: Claire.Pagetti@onera.fr
https://www.onera.fr/fr/staff/claire-pagetti

AI for Air Traffic Management and Large Scale Urban Mobility
Mobility Management: Daniel Delahaye (ENAC)
This project has two related parts: 1) investigating automation in air traffic management, 2) applying new AI algorithms to UAV large scale trajectory planning. We expect to develop new adaptive self- organization algorithms in order to adapt, in real time, the UAV demand to the actual capacity of the airspace. The main difficulty of this research is linked to the certificability of the proposed solutions in order to be implemented with real air vehicles.
Email: daniel@recherche.enac.fr
ferrer”>https://www.math.univ-toulouse.fr/~gamboa/
IPC - INTEGRATIVE PROGRAM C: ASSISTANT FOR DESIGN, DECISION, AND OPTIMIZED INDUSTRY PROCESSES

This IP will develop new Al methods to aid human decisions. This program will design advanced Al assistants to increase the performance of design, decision and industrial production related activities. This will lead to the design of cognitive assistants with advanced dialogue and interaction skills, the monitoring of complex systems in order to model their behaviour, predict their evolution, and anticipate corrective actions, and the design of autonomous mobile robots with the ability to interact with humans, cognitively and physically, to perform complex tasks in a collaborative manner.

Reverse-engineering the brain to build machines that can see and interpret the visual world as well as humans do.

Reverse-engineering the brain: Thomas Serre (External, Brown University, US)
This project will develop ML algorithms that can process visual data in ways that are closer to what humans are capable of. That is, such systems will be robust and reliable though perhaps lose some of the performance of pure ML systems for certain tasks.

Deep Learning with semantic, cognitive and biological constraints

Neuro-inspired Deep Learning: Rufin van Rullen (DR CNRS, CerCo)
This project brings experts from several disciplines in a multi-pronged approach to cognitive/bio-inspired models. It will study multimodal interactions in human brains as a source for more robust, less data demanding ML. AI algorithms from distributed intelligence will also be developed.
Email: rufin.vanrullen@cnrs.fr
http://www.cerco.ups-tlse.fr/~rufin/

Neuro-adaptive Technology based Mixed-initiative to enhance Man-Machine Teams

Neuroergonomics PI: Frédéric Dehais (ISAE)
The chair intends to study flexible mixed-initiative planning and execution paradigm involving humans interacting with artificial agents. The implementation of such an interaction will presuppose to develop passive Brain Computer Interface (pBCI) also known as Neuro-adaptive technology to sense human performance. The project will especially focus on the design of Neuroadaptive technology dedicated to measure multiple users brains while interacting with each other and artificial agents.
Email: Frederic.Dehais@isae-supaero.fr
http://personnel.isae-supaero.fr/frederic-dehais/

Human robot interactions for cobot-industry applications.

Human Robot Interaction PI: R. Alami (DR CNRS, LAAS)
This project will integrate AI with a robotics research program for cognitive and interactive robot partners to develop autonomous teammate robots working with humans, cognitive and interactive assistants for frail people, and highly adaptive service robots.
Other essential aspects that will be investigated include: 1) a principled and long-term multi-disciplinary collaborative research with philosophers, development psychologists, ergonomists, and 2) project-based deployment of AI-enabled robotic systems with potential users and mainly therapists as well as manufacturing and service industry.
Email: rachid.alami@laas.fr
http://homepages.laas.fr/rachid/

Motion Generation for Complex Robots using Model-Based Optimization and Motion Learning

Motion PI: Nicolas Mansard (DR CNRS, LAAS)
The research proposed in this chair aims to generate complex motions for robots in real time by: 1) relying on massive off-line caching of pre-computed optimal motions that are 2) recovered and adapted online to new situations with real-time tractable model predictive control and where 3) all available sensor modalities are exploited for feedback control going beyond the mere state of the robot for more robust behaviors. The goal is to develop a unified yet tractable approach to motion generation for complex robots with arms and legs.
Email: Nicolas.Mansard@laas.fr
http://homepages.laas.fr/nmansard/

Knowledge compilation techniques for reducing complexity of algorithms for solving problems with uncertainty and preferences

Knowledge Compilation PI: Hélène Fargier (DR CNRS, IRIT)
This project will investigate methods for compiling computations needed to solve combinatorial decision problems with preferences and uncertainty (typically above NP) transforming them into a simpler approximation. As examples of such methods, one can exploit the preference structure, reducing the size of the problem by pruning away undesirable options, or by fusing options that are similar enough or by simplifying the description language options distinguishable in a more complex language are unified.
Email: fargier@irit.fr
http://www.irit.fr/~Helene.Fargier/

Synergistic transformations in model based and data based diagnosis

Diagnosis: Louise Travé-Massuyes (DR CNRS, LAAS)
This project will synergistically analyze transformations from model based diagnosis to exhibit fault indicators and data transformations from data based diagnosis methods. The first objective is to highlight and understand the correspondences that may exist between them and how they could complement each other. The second objective is to be able to abstract up data configurations and map them to models suitable for diagnosis reasoning.
Email: louise@laas.fr
http://homepages.laas.fr/louise

Design using intuition and logic

Design using intuition and logic PI: T. Schiex (DR Inra, MIAT)
Associated researchers: Sophie Barbe (INSA), David Simoncini (IRIT)

With the target of providing an enhanced toolkit for designing complex systems by combining logic and intuition, we will extend the theory, algorithms and implementation underlying our guaranteed graphical model solver (toulbar2) to address more complex reasoning problems for NP-hard problem solving from optimization, quantified reasoning or counting. To guide reasoning so that it both finds solutions faster and is able to take into account information extracted from data by Machine Learning, we will integrate ML technology inside our algorithms and models to also solve multi-criteria problems that account both for physical laws represented as logical rules or criteria as well as ML extracted information. The design of complex molecular systems such as proteins will be used as our main target throughout the project, to validate and enhance the visibility of our progresses.
Email: thomas.schiex@inra.fr
http://www.inra.fr/mia/T/schiex

Starting operations this autumn, ANITI will fund upwards of 40 PhD and 30 Post Doc positions.

Successful candidates will have a unique opportunity of contributing to the ambitious research agenda of ANITI, and will be given excellent conditions for the development of their research skills, in terms of working conditions and laboratory facilities, in one the 24 chairs funded.

POST-DOC POSITIONS, PhD POSITIONS

The list of Chairs offering open PhD and Post-Doc positions with their contact information is available here

Applications should be sent by email to:
aniti-phd@univ-toulouse.fr for PhD applications
aniti-postdoc@univ-toulouse.fr for Post-Doc applications
Formal applications should include detailed CV, a motivation letter and transcripts of bachelors’ degrees. Samples of published research by the candidate and reference letters will be a plus.

Events

3IA : Toulouse présélectionnée pour son projet d'institut interdisciplinaire

Le projet ANITI, Artificial and Natural Intelligence Toulouse Institute, porté par l'Université fédérale Toulouse Midi-Pyrénées été présélectionné par un jury international, le 6 novembre 2018.
Lire la suite

Transports autonomes · Séduits par le projet Aniti, les industriels s’engagent

L’IRT Saint Exupéry et l’Université fédérale Toulouse Midi-Pyrénées ont signé un accord de partenariat renforçant l’émulation scientifique en intelligence artificielle qui lie les entreprises du domaine du transport/mobilité avec les laboratoires de recherche publics toulousains.
Lire la suite

3IA · Le projet ANITI déposé à l'ANR

Le dossier complet du projet ANITI : Artificial and Natural Intelligence Toulouse Institute, pour accueillir à Toulouse en 2019 l’un des instituts interdisciplinaires dédiés à l'intelligence artificielle (3IA), a été déposé mardi 19 février 2019 auprès de l’Agence nationale de la recherche (ANR) ...
Lire la suite

ANITI, le projet d'Institut d'Intelligence Artificielle toulousain est selectionné !

Le projet toulousain ANITI porté par l'Université fédérale Midi-Pyrénées est sélectionné en tant qu’institut 3IA.
Lire la suite

Présentation d'ANITI au salon du Bourget

L'institut interdisciplinaire d’intelligence artificielle de Toulouse - ANITI - présenté au salon du Bourget à Paris.
Lire la suite

Plate-Forme Intelligence Artificielle PFIA 2019

L’objectif de la Plate-Forme Intelligence Artificielle (PFIA) est de réunir chercheurs, industriels et étudiants autour de conférences et d’ateliers consacrés à l’Intelligence Artificielle (IA).
Lire la suite

Événement partenaires d’ANITI

Lancement officiel d'ANITI à l'Université fédérale Toulouse Midi-Pyrénes, en présence des partenaires socio-économiques et académiques.
Lire la suite

Journées Scientifiques ANITI

Deux journées permettant aux scientifiques impliqués dans les chaires, de se présenter et d'établir une proposition de programme scientifique commun.
Lire la suite

ANITI - Forum industriel

ANITI organise une rencontre entre chercheurs académiques et experts industriels. La journée est organisée autour de tutoriels donnés par les chercheurs et des sessions de résolution de problème associés.
Lire la suite

Robotics in the 3IA Prairie and Aniti

Robotics is one of the ultimate challenge for artificial intelligence. It encompasses many open challenges: computer vision, human-machine interface, reinforcement learning, ...
Lire la suite

FUTURAPOLIS à Toulouse

Deux jours de débats, de découvertes et d’activités ludiques sur l’innovation et le High-Tech en plein coeur de Toulouse.
Lire la suite
Université fédérale
Toulouse Midi-Pyrénées

41 allées Jules Guesde – CS 61321
31013 Toulouse – CEDEX 6
Tel. + 33 (0)5 61 14 80 10

contact.aniti@univ-toulouse.fr
Come to the University

ANITI, NEW MAJOR HUB FOR ARTIFICIAL INTELLIGENCE (EN)

download pdf (1,9Mo)