RT @EliJalbout: I'll be presenting this work next week @ieeeras #ICRA40 🚀 Looking forward to insightful discussions and (re)connecting with….

RT @EliJalbout: 🧵 Let's take a closer look at how the action space impacts #RobotLearning and #SimToReal transfer. 🤖🚀📈🔧 #AI . Blog: https:….

RT @EliJalbout: Ever wondered what action space to choose for your manipulation task? or which action space is most suited for sim-to-real….

#NeurIPS paper on imitation learning: our method AIME can learn from observing videos of Expert behaviour. Blog post at Poster #1501 on Wednesday 5–7 pm, presented by the first author Xingyuan Zhang.

RT @ThierryBreton: New day, same trilogue! ⏰. Entering the 20th continuous hour of #AIAct negotiations.

New lab paper accepted -- showing where we are taking latent state-space models next.

RT @etamiAI: Join etami‘s public online event on Monday, January 23, at 3pm CET and learn more about us and our mission to translate ethica….

Great results by Atanas, Justin, and the SLAM team.

RT @etamiAI: We are happy to announce etami’s transition into a @BDVA_eu Task Force! etami is open to BDVA members who want to actively sha….

Our SLAM work is progressing well thanks to Baris Kayalibay, Atanas Mirchev, and @usuallyuseless.

RT @etamiAI: Next online #etami lecture on May 31, 12:00–13:00 CET: Charlotte Ducuing from @CiTiP_KULeuven. speaks on data regulation. Free….

RT @etamiAI: #etami & @CiTiP_KULeuven are proud to present session #2 of the etami® lecture series. Free tickets for our session on 25th….

RT @etamiAI: Introducing #etami® – Ethical and Trustworthy Artificial and Machine Intelligence. Join our mission towards Artificial Intelli….

Help spread the word and support! As Fellow of this 42 school I know @maxsenges and his team as amongst the most driven and conscientious people around. Here's their most important call yet.

9 days left to register for the #Robothon – The Grand Challenge. I'm happy to again be part of the jury to judge your robotic solutions e-waste disassembly. Apply at

Still free seats free for this Monday's etami first lecture – on #explainableAI. Register at

RT @etamiAI: The #etami lecture series starts soon, with our partner @KU_Leuven (Centre for IT & IP Law). Join our.public online lectures o….

Latent matters! In our new blog post we explain a method published at NeurIPS to make your deep state-space models work, and prevent training to get stuck or predictions to fail.

Sequential latent-variable models don't always work well, for two reasons: (1) the Gaussian prior is too weak; (2) the encoder/decoder represent too much of the system dynamics. Our latest NeurIPS paper solves this. This Thursday, 12:30am PST / 21:30 CET.

We have open-sourced our drone hardware at Our drones had to learn to fly using Deep Variational Bayes Filters. Their motor babbling sometimes caused them to hit the wall -- so we needed a design that could absorb the impacts.