The Conference on Robot Learning (CoRL) is an annual international conference specialised in the intersection of robotics and machine learning. The fifth edition took place last week in London and virtually around the globe. Apart from the novelty of being a hybrid conference, this year the focus was put on openness. OpenReview was used for the peer review process, meaning that the reviewers’ comments and replies from the authors are public, for anyone to see. The research community suggests that open review could encourage mutual trust, respect, and openness to criticism, enable constructive and efficient quality assurance, increase transparency and accountability, facilitate wider, and more inclusive discussion, give reviewers recognition and make reviews citable [1]. You can access all CoRL 2021 papers and their corresponding reviews here. In addition, you may want to listen to all presentations, available in the conference YouTube channel.
In this post we bring you a glimpse of the conference through the most popular tweets written last week. Cool robot demos, short and sweet explanation of papers and award finalists to look forward to next year’s edition in New Zealand. Enjoy!
Robots, robots, robots!
Here’s our fourth and final #CoRL2021 paper, on multi-stage imitation learning: https://t.co/22JqLMMp4X
We’re doing a live demo of this right now, see below! This sequence was trained with a single demonstration, and generalises across poses / distractors.
With @normandipalo. pic.twitter.com/ENEzKvAlRm
— Edward Johns (@Ed__Johns) November 11, 2021
ANYmal trotting around at #CoRL2021 pic.twitter.com/59ql6sEJEx
— Kai Arulkumaran (@kaixhin) November 9, 2021
When I am human – I like dogs a lot. But it turns out that when I am a robot – I am a bit timid around quadrupeds. Thanks to the wonderful team of #CoRL2021 who organized telepresence sessions (@Ed__Johns, @vitalisvos19, Binbin Xu, and others). This was a very curious experience! pic.twitter.com/mMEdqUq6yv
— Rika Antonova (@contactrika) November 11, 2021
We will be doing a live demo of iMAP at #corl2021, sessions 1 and 8 Monday and Thursday, come if you're around!https://t.co/Tagk4jocbe@liu_shikun, @joeaortiz, @AjdDavison pic.twitter.com/XTUYrkOWnO
— Edgar Sucar (@SucarEdgar) November 8, 2021
Papers and presentations
Excited to present our work at #CoRL2021 this week. Our paper shows a promising approach to achieve sim2real transfer for perception. w/ @ToyotaResearch @berkeley_ai (1/7) https://t.co/MAbYaXaPrY pic.twitter.com/scfv4DtOSJ
— Michael laskey (@Michaellaskey7) November 8, 2021
Ever wondered how to tune your hyperparameters while training RL agents? w/o running thousands of experiments in parallel? And even combine them?
Check out our work @ #CoRL2021 on training mixture agents which combines components with diverse architectures, distributions, etc pic.twitter.com/4tsqWXjf79
— Markus Wulfmeier (@markus_with_k) November 9, 2021
Proud to share "LILA: Language-Informed Latent Actions," our paper at #CoRL2021.
How can we build assistive controllers by fusing language & shared autonomy?
Jointly authored w/ @megha_byte, with my advisors @percyliang & @DorsaSadigh.
: (1 / 10) pic.twitter.com/K5vdUO8R6q
— Siddharth Karamcheti (@siddkaramcheti) November 8, 2021
: In daily life after passing the paper exam for a permit, a newbie can drive the car for real under an expert's monitoring. This guardian mechanism is crucial for exploratory but safe learning. See our #CoRL2021 work https://t.co/OegFMwj9JS
Poster session today at Session 8 pic.twitter.com/IlKkZhk8UD— Bolei Zhou (@zhoubolei) November 11, 2021
I've just finished my talk in the Blue Sky oral session at #CoRL2021. Great to have a session which allows for unconventional opinions to be heard in this way. Looking forward to the next two!
Here's my paper, Back to Reality for Imitation Learning: https://t.co/TZLj77zGWV. pic.twitter.com/NnwZNaSsX5
— Edward Johns (@Ed__Johns) November 9, 2021
This week at #CoRL2021, our team explores how data efficient off-policy RL, on-board reward generation, and vision can teach walking and wall avoidance to small humanoid robots without laboratory instrumentation: https://t.co/ZZo9hJXXO5 1/2 pic.twitter.com/4qWqn1hefF
— DeepMind (@DeepMind) November 10, 2021
Awards
Congratulations again to the #CoRL #BestSystemPaper Winner, "FlingBot: The Unreasonable Effectiveness of Dynamic Manipulation for Cloth Unfolding", by Huy Ha and Shuran Song
Read it here: https://t.co/en8ic29gPM#robot #learning #robotics #award pic.twitter.com/P9ZW884dDn— Conference on Robot Learning (@corl_conf) November 12, 2021
Congratulations again to the #CoRL #BestPaper Winner, "A System for General In-Hand Object Re-Orientation", by Tao Chen, Jie Xu and Pulkit Agrawal
Read it here: https://t.co/QeIzAW1fng#robot #learning #robotics #award pic.twitter.com/niHDCArbmq— Conference on Robot Learning (@corl_conf) November 12, 2021
Congratulations to #CoRL2021 best paper finalist, "Robot Reinforcement Learning on the Constraint Manifold", Puze Liu, Davide Tateo, Haitham Bou Ammar, Jan Peters.https://t.co/SdssOu3akD#robotics #learning #award #research pic.twitter.com/Lpgn2XjEhY
— Conference on Robot Learning (@corl_conf) November 8, 2021
Congratulations to #CoRL2021 best paper finalist, "Learning Off-Policy with Online Planning", Harshit Sikchi, Wenxuan Zhou, David Held.https://t.co/7xUFyMlyQU#robotics #learning #award #research pic.twitter.com/D5u4NdDDST
— Conference on Robot Learning (@corl_conf) November 8, 2021
Congratulations to #CoRL2021 best paper finalist, "XIRL: Cross-embodiment Inverse Reinforcement Learning", Kevin Zakka, Andy Zeng, Pete Florence, Jonathan Tompson, Jeannette Bohg, Debidatta Dwibedi.https://t.co/Z4qVMiuhgS#robotics #learning #award #research pic.twitter.com/k2KSkv67vd
— Conference on Robot Learning (@corl_conf) November 8, 2021
Congratulations to #CoRL2021 best systems paper finalist, "SORNet: Spatial Object-Centric Representations for Sequential Manipulation", Wentao Yuan, Chris Paxton, Karthik Desingh, Dieter Fox.https://t.co/en8ic29gPM#robotics #learning #award #research pic.twitter.com/Hfn1CiZNP8
— Conference on Robot Learning (@corl_conf) November 8, 2021
Congratulations to #CoRL2021 best systems paper finalist, "Fast and Efficient Locomotion via Learned Gait Transitions", Yuxiang Yang, Tingnan Zhang, Erwin Coumans, Jie Tan, Byron Boots.https://t.co/HbrCOVZOpz#robotics #learning #award #research pic.twitter.com/T5Ev31PgPZ
— Conference on Robot Learning (@corl_conf) November 8, 2021
