SystemX is organising the 2nd International Workshop on Artificial Intelligence and Augmented Engineering (AIAE’22), at Inria (Palaiseau), on December 1st, 2022.

About the event

The Deep Learning revolution over the last decade has progressively invaded all fields of digital science and is gradually establishing itself as a complement to physical and knowledge-based models at all stages of modeling, simulation, optimization and control of complex systems.
Following the successful first edition in 2021, starting as an echo of the IA2 Program at IRT SystemX, the second workshop, co-organized with Systematic and the DATAIA Institute, will feature two international keynote presentations, and will allow researchers to present their most recent works. Contributed presentations are solicited, covering various topics in the field, including but not limited to the following:

  • Handling and explaining the massive output data of heavy numerical simulations
  • Accelerating numerical simulations with Deep / Machine Learning
  • Improving the accuracy or the robustness of simulations with Machine Learning
  • Learning to solve ODEs and PDEs
  • Discovering mechanistic/behavioral models from data
  • Incorporating physical constraints in Deep Learning
  • Providing support in augmented decision making for complex systems
  • New Human-Computer Interactions
  • Symbolic and Numerical AI hybridization
  • Natural language processing techniques based on IA
  • Hybridization of deep learning with symbolic artificial intelligence
  • Knowledge extraction and reasoning techniques from heterogeneous data
  • Ontology and knowledge graph alignment
  • Stream processing and reasoning (video and textual data)

Scientific committee

  • Patrice Aknin (IRT SystemX)
  • Faicel Chamroukhi (IRT SystemX)
  • Eric Duceau (ENPC)
  • Céline Hudelot (CentraleSupélec)
  • Mostepha Khouadjia (IRT SystemX)
  • Juliette Mattioli (Thales Group)
  • Marc Schoenauer (Inria)
  • Mouadh Yagoubi (IRT SystemX)


8:30am – 9:00am | Opening reception

9:00am – 9:30am | Opening session
Marc Schoenauer, Research Director Inria Saclay – Ile-de-France, Scientific Coordinator, IA2 program – IRT SystemX

9:30am – 10:30am | Keynote #1:
Michael Witbrock, Professor of Computer Science, The University of Auckland, New-Zealand
Can AI Reasoning be Probably Approximately Correct Enough for Engineering?
Machine Learning-Based and Symbolic AI Systems have been divided by a chasm; the learned systems are general, but often inaccurate; the symbolic systems are precise but almost impossible to apply. Recent research has begun to focus on systems that have broad coverage, but that learn to reason more accurately. In this talk, we shall consider how accurate these systems need to be, and how broad, and will consider a couple of cases where deep learning and symbolic systems are beginning to come together with the promise of reasoning well enough – forming systems whose reasoning is (quantifiably) probably approximately correct.

10:30am – 11:00am | Coffee break and posters

11:00am – 12:30am | Parallel sessions
Session 1: Machine Learning and Physics : PINNs approaches
Chair: Marc Schoenauer – Room Sophie Germain Lecture Hall

  • Frederic Barbaresco (Thales) Lie Groups Machine Learning & Foliation Structures of Thermodynamics-Informed Neural Networks
  • Mathieu Riou, Teodora Petrisor (Thales)  Physics and Geometry Informed Neural Networks for Industrial Applications
  • Antoine Benady, Emmanuel Baranger, Ludovic Chamoin (LMPS – ENS Paris-Saclay)  Physics-informed neural networks derived from a mCRE functional for constitutive modelling
  • Thi Nguyen Khoa Nguyen (Centre Borelli – ENS Paris-Saclay) Physicsinformed neural networks for non-Newtonian fluid thermo-mechanical problems: An application to rubber calendering process

Session 2 : Knowledge based model meets Machine Learning
Chair: Céline Hudelot Room Gilles Kahn Hall

  • Juliette Mattioli, Claire Laudy, Pierre-Olivier Robic, Hugo Guillermo Chalé Gongora (Thales)  A hybrid AI knowledge graph approach to support Body of Knowledge design
  • Arthur Ledaguenel, Céline Hudelot, Mostepha Khouadjia (IRT SystemX, MICS – CentraleSupélec) Multi-Label Classification with Semantic Projection and Semantic Regularization
  • Etienne Bennequin, Myriam Tami, Antoine Toubhans, Celine Hudelot (Sicara, MICS – CentraleSupélec) Few-Shot Image Classification Benchmarks are Unrealistic: Build Back Better with Semantic Task Sampling
  • Hannah de Oliveira Plath, Maria Estela de Oliveira Paiva, Danielle Lanzarini Pinto, Paula Costa (Unicamp, Brazil). Hate Speech Detection Against Women in Texts in Brazilian Portuguese. 

12:30pm – 2:00pm | Lunch break and posters of IA2 program of SystemX

2:00pm – 3:00pm |Keynote #2
Petros Koumoutsakos, Professor of Engineering and Applied Sciences, Harvard University, USA
AI/Scientific Computing: Alloys for Flow Modeling and Control
Over the last last thirty years we have experienced more than a billion-fold increase in hardware capabilities and a dizzying pace of acquiring and transmitting massive amounts of data. Artificial Intelligence (AI) has been the beneficiaries of these advances and today it is increasingly embedded in technologies that touch every aspect of humanity. However along with the abundance of promise there is an ever increasing amount of hype, in particular regarding the capabilities of learning algorithms to model, predict and control complex fluid mechanics problems. In this talk I would offer a perspective on forming alloys of AI and numerical methods for the prediction and control of complex flow systems. I will present novel algorithms for learning the Effective Dynamics (LED) of complex flows and a fusion of multi- agent reinforcement learning and scientific computing (SciMARL) for modeling and control of complex flow-structure interactions. I will juxtapose successes and failures and argue that the proper fusion of fluid mechanics knowledge and AI expertise are essential to advance scientific frontiers.

3:00pm – 5:20pm | Parallel sessions
Session 3: Machine Learning and Physics: ODEs/PDEs
Chair: Mouadh Yagoubi Room Sophie Germain Lecture Hall

3:00pm – 3:40pm

  • Jean-Pierre Merlet (Inria) Trying AI for solving kinematic equations in robotics
  • Gilles Kluth, Loic Lepareur (CEA DAM) Deep learning to speed-up and enhance inertial confinement fusion calculations

3:40pm – 4:00pm| Coffee break and posters

4:00pm – 5:20 pm

  • Pierre Benjamin (Airbus) Strong coupling between boundary element method and a learning model
  • Bilel Bensaid, Gael Poette, Rodolphe Turpault (Institut de Mathématique de Bordeaux, CEA-CESTA) Deterministic neural networks optimization from a continuous and energy point of view
  • Hannah Plath, Thiago Petrilli Dardis Maffei, Alexandre Allauzen (ESPCI-PSL, MILES team – Lamsade – Univ. Dauphine).  Experimental study of Neural ODE training with adaptive solver for dynamical systems modeling
  • Daniele Noto, Sergio Chibbaro, Alexandre Allauzen (Sorbonne Université, LISN – Université Paris-Saclay, Lamsade – Univ. Dauphine). Efficient training method for a shell model of turbulence

Session 4: Machine Learning and Physics: Learning methodologies
Chair: Faïcel Chamroukhi Room Gilles Kahn Hall
3:00pm – 3:40pm

  • Alban Puech, Mohamed Alami Chehboune, Walter Telsnig, Anatoliy Zabrovskiy, Martin Göldner, Jesse Read (LIX – Polytechnique, IRT SystemX, DEIF Wind Power Technology Austria) An Improved Yaw Control Algorithm for Wind Turbines via Reinforcement Learning
  • Vincent Le Guen (EDF) Deep augmented physical models: application to model-based reinforcement learning

3:40pm – 4:00pm| Coffee break and posters

4:00pm – 5:20 pm

  • Paul Boniol (Lipade – Univ. Paris Cité) Detection of Anomalies and Identification of their Precursors in Large Data Series Collections
  • Antoine de Mathelin, François Deheeger, Mathilde Mougeot, Nicolas Vayatis (ENS Paris-Saclay – Michelin). From theoretical to practical transfer learning: the ADAPT library
  • Thiago Petrilli Maffei Dardis, David Benhaiem, Robbie Radjagobalou, Jérémie-Luc Sanchez (Baalbek Management, Chimie ParisTech) Optimization of photochemical reaction conditions using Bayesian optimization and machine learning

5:20pm – 5:30pm |Hub DSAI actions of Systematic cluster 
Juliette Mattioli, Senior Expert in Artificial Intelligence, Thales

5:30pm | Conclusion
Paul Labrogère, CEO, IRT SystemX

5:40pm | Closing cocktaill


Registration is now closed. For any additional registration, please send an e-mail to

Practical information

Inria – Amphithéâtre Sophie Germain
1rue Honoré d’Estienne d’Orves, 91120 Palaiseau
Access map

Inscrivez-vous à la newsletter de l'IRT SystemX

et recevez chaque mois les dernières actualités de l'institut :