Objectives
Inria’s RBA center and IRISA are organizing an internal event on AI in our teams, from contributors to the field to potential users. Over 1.5 days, general presentations (scope of AI, fundamentals of machine learning, deep learning), presentations from volunteering teams in our labs and focused scientific posters will draw a panorama of our work on AI to foster scientific exchanges, mean factorization and to open new perspectives and scientific challenges to address. The general idea of this informal event is to better define what we mean by AI, better undertand what related skills are present in our labs, what research directions are pursued today and could be pursued tomorrow. All teams, regardless of their affiliations and area of expertise, are very welcome.Program
Thursday, April 19
| 9h00 | 9h30 | Coffee and introduction |
| 9h30 | 10h15 | Bertrand Braunschweig. What's AI? |
| 10h15 |
11h00 |
Liva Ralaivola. A general introduction to machine learning. |
| 11h00 |
11h15 |
Break |
| 11h15 |
11h35 |
Arnaud Martin (DRUID). Managing and qualifying uncertain, participative, massive and interlinked data |
| 11h35 |
11h55 |
Sébastien Ferré and Peggy Cellier (SEMLIS): Knowledge graphs, NLP and information extraction |
| 11h55 |
12h15 |
François Coste (DYLISS): Knowledge acquisition on functions of biological sequences and systems |
| 12h15 |
12h35 |
Camille Maumet (VISAGES): Towards large-scale brain imaging studies: How to deal with analytic variability? |
| 12h35 |
13h45 |
Buffet lunch |
| 13h45 |
14h30 |
Stéphane Canu. Deep learning explained. |
| 14h30 |
14h50 |
Laetitia Chapel (OBELIX): AI and earth observation |
| 14h50 |
15h10 |
Olivier Le Meur (PERCEPT): Visual perception and deep learning |
| 15h10 |
15h30 |
Guillaume Gravier (LINKMEDIA): Understanding media and gaining knowledge |
| 15h30 |
16h30 |
Posters & Coffee |
| 16h30 |
16h50 |
Erven Rohou (PACAP): AI for processors and processors for AI |
| 16h50 |
17h10 |
Olivier Sentieys (CAIRN): Approximate deep learning accelerators |
| 17h10 |
17h30 |
Rémi Gribonval (PANAMA): Deep networks: engineered, trained or randomized? |
Friday, April 20
| 9h20 |
9h40 |
Paolo Robuffo Giordano (RAINBOW): Robots able to navigate and/or interact in complex environments (bonus on deep learning for visual servoing by Éric Marchand) |
| 9h40 |
10h00 |
Gerardo Rubino (DIONYSOS): random neural networks and extensions - applications in computer networks |
| 10h00 | 10h20 | Erwan Faou (MINGUS): Machine learning vs. modelling in scientific computing |
| 10h20 | 10h40 | Mathieu Acher (DIVERSE): Software variability and artificial intelligence |
| 10h40 | 11h00 | Break |
| 11h00 |
11h20 |
Alexandre Termier (LACODAM): Making sense from large quantities of data |
| 11h20 |
11h40 |
Bertrand Couasnon (INTUIDOC): Grammar, incremental learning and deep neural networks |
| 11h40 |
12h00 |
Gwénolé Lecorvé (EXPRESSION):
Expressive communicating machines |
| 12h00 |
12h15 |
Concluding remarks |
Posters
In no particular order:Kevin Bannier. Deep learning methods for saliency prediction on webpages and comics.
Léo Maczyta. Détection de la saillance dynamique dans des vidéos par apprentissage profond.
Paul Temple. Multimorphic Testing for Machine Learning based Software Systems.
Thierry Dumas. Deep transform for image compression: Can it be universal?
Aurélien Lamercerie. Analyse formelle d’exigences en langue naturelle pour la conception de systèmes cyber-physiques
Emmanuel Moebel. Localisation de macro-molécules dans des images 3D de cryo-tomographie électronique.
Himalaya Jain. SUBIC: a supervised, structured binary code for image search.
Adrien Luxey. Sprinkler: a probabilistic dissemination protocol to provide fluid user interaction in multi-device ecosystems.
Raphaël Truffet. Optimal acquisition design for sparse reconstruction using dictionary learning in diffusion magnetic resonance imaging
Ian Jeantet. Understanding the evolution of science.
Cédric Maigrot. Detecting fake news in social networks.
Anne-Lyse Minard. Deep learning for opinion mining.
Clément Dalloux.Towards deep understanding of clinical records.
Kwon-Young Choi. Unsupervised Segmentation of Music Symbols using a Detector in Adversarial Training.
Olivier Ruas. Fingerprinting big data.
Corentin Hardy. Evaluation of Random Neural Layers in Deep Neural Networks.
Registration and practical information
When: April 19-20, 2018Where: Inria's conference center. There will also be a visio setup for distant sites to attend. Details on the visio to follow.
Registration: Registration is now closed. You can still attend but don't complain if there's not enough food or coffee!
Useful links
- Inria’s whitebook on AI
- a cartography of French labs contributing to AI, from Journal du Net
- the “France IA” report handled to the goverment in 2017 (in French, sorry)
- the report of ALLISTENE regarding the creation of an AI infrastructure in France (in French again, sorry)
- When computer decides: European recommendations on machine-learned automated decision making - a 2018 report of Informatics Europe & EUACM
- Le fameux rapport Villani : Donner un sens à l'intelligence articifielle pour une stratégie nationale et européene
Contact
Workshop coordination by Elisa and Guig. Please contact us if you have any question or suggestion. Practical organization and logistics by Aurélie, with help from IRISA and Inria communication services.A definition of AI
Wanna know what AI is about? Check the “official” definition from the Villani mission on AI!! And feel free to comment, approve, criticize, etc.
Registration now opened
Registration to the AI days is now opened. It’s free, opened to everyone affiliated to a research team within Inria RBA or IRISA, and simple: put your name on the following poll and that’s it. Registration is however mandatory to help us adjust the number of person at the pauses…
Les annonces du gouvernement
Après la remise du rapport Villani au gouvernement, ce dernier fait ses annonces. Retrouvez une synthèse de ces annonces dans un article du monde.fr en date du 29 mars 2018. Pour tout comprendre de ces annonces, il est intéressant de consulter le rapport Villani et de consulter le site correspondant…
Call for demos and poster presentations
In addition to general presentations and team presentations, the event features technical poster presentations, where real hard-core science can be presented. We also welcome project presentations or any other poster presentation that will bring interesting information regarding AI activities in our labs. Demos are obviously also welcome. All these presentations…
Bertrand Braunschweig to make the introduction to AI
We’re happy and proud to announce that Bertrand Braunschweig accepted to give a general presentation on AI in general, following the cataloging work he did for Inria. Do not hesitate to check the white book online!
Save the dates!
The workshop will take place on April 19 and 20, 2018, at the Inria conference center. Thank you for all those who took the time to doodle. We booked the dates where most people were available. Waiting forward to seeing you at the workshop. And stay tuned for the program!
