AI Safety, Explainability and Robustness Team at CEA. Tools: - CAISAR: https://xitter.cf/t.co/m9KJaC0FkU - PyRAT: https://xitter.cf/t.co/zlIKcDfxk5 - Colibri*: https://xitter.cf/t.co/dqirmPTGNx
Paris, France
Joined February 2023
- Tweets17
- Following43
- Followers29
- Likes6
Due to a combination of factors, the main one being the huge exile of academics (our main audience) out of X, especially in the field of AI/ML, we no longer feel it is suited for our continued presence.
You can still find us on Mastodon: fosstodon.org/@aiser_team
We are proud to announce that our analyzer PyRAT scored second during the Fifth International Verification of Neural Networks Competitions, which was co-hosted with the conference CAV and the symposium SAIV! 🎉
For more info, see: caisar-platform.com/2024/07/…
June 19th was the 34th birthday of the abolition of apartheid laws in South Africa. To celebrate this, we released the version 2.0 of CAISAR. You can read more about it here: caisar-platform.com/2024/06/…
The release is available through our public repository, on dockerhub and opam!
Last Friday was the 68th birthday of the Flag of Europe. To celebrate this, we released the version 1.0 of CAISAR.
You can read more about it here: caisar-platform.com/2023/12/…
The release is available through our public repository, on dockerhub and soon on opam!
Last week we presented CAISAR at the workshop on Safe and Robust Machine Learning.
We want to thank the Lab for AI Verification (laiv.uk/) for having had us, as well as for the very interesting conversations.
#ArtificialIntelligence #AI #MachineLearning #AISafety
We have several Master 2 internship positions, covering a rather broad range of topics.
The offers are available on our website: caisar-platform.com/position…
Feel free to largely disseminate these offers to your colleagues and students, or to contact us if you wish to apply!
We had a great time yesterday presenting our 2 papers at @WAISE_AIsafety about #ArtificialIntelligence monitoring and #AISafety! Thank you again for having us.
Now, time to enjoy more great talks at @safecomp2023 😄
#MachineLearning #TrustworthyAI #SAFECOMP2023
ALT Researcher presenting their work and result to other people.
ALT Researcher presenting their work and result to other people.
How can #FormalMethods help us improve the safety, robustness and trustworthiness of #ArtificialIntelligence and #MachineLearning systems? Our tool, PyRAT, can help on that front. Feel free to check it out here for more information: pyrat-analyzer.com/
🤝👾L’adoption massive de l’intelligence artificielle ? Une question de sûreté et de robustesse. Notre outil PyRAT, basé sur les méthodes formelles, permet de vérifier que le comportement d’une #IA n’est pas modifié par de légères perturbations... Découvrez-le en vidéo ! 👇
The efficiency of Symbolic #AI applied to #softwaresecurity & #safety is undeniable! Our tool COLIBRI got first place in 8 categories of the SMT-COMP! Congratulations to Bruno Marre, the lead developer, as well as all those who helped get COLIBRI to where it is today 💪
AISER Team retweeted
[#ListTechDays] Merci à tous nos partenaires d'avoir partagé nos dernières innovations et nourri nos échanges.
🎬Retour en vidéo sur cette belle journée ! #France2030 @Safran @Siemens @DGAC @iledefrance @Reseau_Carnot @CEA_Officiel @SGPI_avenir #Numérique #Tech #Innovation
Yesterday was the 234th birthday of the Tennis Court Oath. To celebrate this, we released the version 0.2 of CAISAR.
You can read more about it here: caisar-platform.github.io/we…
The release is available through our public repository, on dockerhub and soon on opam!
AISER Team retweeted
[#SIAE 🚀] Retrouvez les chercheurs du @CEA_Officiel au salon du Bourget, avec leurs technologies innovantes !
En particulier, A. Lemesle de la @caisar_team présente la technologie #PyRAT pour la #robustesse et la #sûreté des #IA.
👉stand B42, hall4
@CEA_List @CEA_Liten @CEA_Leti
AISER Team retweeted
[#IA] Souveraineté, frugalité et confiance : le collectif #DeepGreen, piloté par le @CEA_Officiel, lance la première plateforme d'IA embarquée française en open source.
🗞 Voir le communiqué de presse bit.ly/41G7Qma
We have a stand at @ETAPSconf 2023 alongside our colleagues @frama_c and @BinsecTool. Feel free to come and talk about AI safety and trustworthiness with us!
#ArtificialIntelligence #AI #MachineLearning #TrustworthyAI #AISafety
AISER Team retweeted
[#ACTU] Programme @ConfianceAI |🧠Peut-on faire confiance aux modèles d'#IA ?
Le @CEA_List propose AIMOS : un outil de test automatisé qui assure la robustesse des systèmes d'IA face à des perturbations et valide leur bon fonctionnement.
@caisar_team
list.cea.fr/fr/19-avril-2023…
Issue 11 of @ActuIAFr magazine is out and features two of our tools (PyRAT end AIMOS)!
#ArtificialIntelligence #MachineLearning
actuia.com/actualite/intelli…
