The third edition of the Reinforcement Learning course at @CS_TUDarmstadt started today, with a very nice introduction by Prof. @GeorgiaChal !
I'm excited to teach some advanced Deep RL methods again this year!
I am happy to announce a workshop on
**Embodiment-Aware Robot Learning (EARL)**
earl.robot-learning.net
at Robotics: Science and Systems Conference #RSS2024 in Delft, Netherlands @RoboticsSciSys
on July 15, 2024
co-organized with @AlapKshirsagar @Jan_R_Peters
In the next @ELLISforEurope Turin Talk, our guest speaker @davide_tateo from TU Darmstadt will present "Robots learning and acting safely in the real world".
Join us on April 4th, 10AM in person at @VandalLabPolito & on Zoom!
See polito.it/ateneo/comunic
@tommasi_tatiana
Co-hosted the Indo 
-
German "Frontiers of Robot Learning" 
workshop at IIT Bombay with Prof. Arpita Sinha, Prof. @Jan_R_Peters, & Prof. Leena Vachhani. Thanks to @INDOGSTC , @iitbombay , @TUDarmstadt and all the participants! More details at igstc.robot-learning.net
- German "Frontiers of Robot Learning" workshop at IIT Bombay with Prof. Arpita Sinha, Prof. @Jan_R_Peters, & Prof. Leena Vachhani. Thanks to @INDOGSTC , @iitbombay , @TUDarmstadt and all the participants! More details at igstc.robot-learning.net
Introducing CrossQ, just published at #ICLR2024! 
CrossQ achieves:
Very fast off-policy Deep RL
with SOTA sample-efficiency in <5% of the gradient steps
without target nets or Q ensembles
Project and Code: adityab.github.io/CrossQ
Joint work w/ @DPalenicek 
CrossQ achieves:
Very fast off-policy Deep RL
with SOTA sample-efficiency in <5% of the gradient steps
without target nets or Q ensembles
Project and Code: adityab.github.io/CrossQ
Joint work w/ @DPalenicek 
Our new flexible-spine quadruped is ready for action!
Quote
Davide Tateo
@davide_tateo
Guess which is one of the first labs to test this amazing platform?
Yes, it's @ias_tudarmstadt!
If you missed our poster @RealAAAI the other day about Parameterized Projected Bellman Operator (PBO), here is a quick rundown:
Instead of optimizing in the space of Q-functions, we learn a "HyperNetwork" that directly maps parameters of the Q-functions
To train the PBO…
Congrats Julen!
Quote
Julen Urain
@thepythagorian
Thrilled to announce my selection as a finalist for the esteemed George Girault Ph.D. award, Europe's highest honor for robotics dissertations.
Congratulations to all the other finalists! 
Happy to share that our paper on Tactile Active Recognition of Textures (TART) got accepted to #ICRA2024!
Super excited that CrossQ got accepted at @iclr_conf! We show how to effectively use #BatchNorm in #RL, yielding SOTA sample efficiency while staying as computationally efficient as SAC!
This is joined work with @aditya_bhatt
#ICLR2024 openreview.net/pdf?id=PczQtTs
We show how to effectively use #BatchNorm in #RL, yielding SOTA sample efficiency while staying as computationally efficient as SAC!
This is joined work with @aditya_bhatt
#ICLR2024 openreview.net/pdf?id=PczQtTs
Very proud for my Ph.D. students that got a paper accepted at @iclr_conf! Amazing job guys! A THREAD on our 4 accepted papers:
Human-Interactive Robot Learning #HIRL workshop is back this year for its 3rd installment at @HRI_Conference'24 on March 15, 2024. We invite #HIRL researchers to submit extended abstracts: sites.google.com/view/hirl24/cfp
LocoMuJoCo won the best poster award at @NeurIPSConf Robot Learning workshop! All credit to the amazing work of @firasalhafez
twitter.com/firasalhafez/s
Join us today at the Robot Learning Workshop during #NeurIPS2023 to checkout #LocoMuJoCo, the first imitation learning benchmark for locomotion. Join us to see our poster and chat about humanoids!
Now even easier to install:
pip install loco-mujoco
github.com/robfiras/loco-
The Air Hockey Challenge 2023 Workshop at @NeurIPSConf was a success! Congratulations to the AiRLIHockey @Idiap_ch team for winning the competition and to all other finalists for making this challenge interesting!
The Robot Air Hockey Challenge is kicking off now in Room 353, and will feature talks from the participants, Chris Atkeson and @svlevine ! 
#NeurIPS2023
#NeurIPS2023
Replying to @ias_tudarmstadt
Quote
Firas Al-Hafez
@firasalhafez
Introducing LocoMuJoCo, the first imitation learning benchmark tailored towards locomotion! It comes with a diverse set of environments ranging from musculoskeletal models to the brand-new @UnitreeRobotics H1 robot, and also many motion capture datasets!
github.com/robfiras/loco-
Replying to @ias_tudarmstadt
Replying to @ias_tudarmstadt
Replying to @ias_tudarmstadt
Quote
Boris Belousov
@_bbelousov
Looking forward to the 1st Workshop on Touch Processing touchprocessing.org/2023/ at #NeurIPS next week!
We will be presenting our latest work on
Tactile Active Texture Recognition With Vision-Based Tactile Sensors
How to identify a given fabric in as few touches as possible?