How is it possible to live in this complex world as it is? What′s life? What′s intelligence?
Postdoc
@Matsuo_lab
@trail_ut
at UTokyo, Robot Learning
#ICRA2024
In the next two weeks, I'll be an (unofficial) local tour guide of
#ICRA2024
since Yokohama is my hometown, where I have lived for over 20 years. So excited! 🚢⛵️⚓️🌉
I'll post the local info of Yokohama in the thread 👉
Our new work is open!!
We developed a "never-give-up" foundation-model-based service robot system.
It has a self-recovery mechanism by prompt updates though reasoning by LLMs and HRI.
Please watch the video how it works!!
Schumidhuberからすんごいタイトルの論文が出てる...
Generative Adversarial Networks are special cases of Artificial Curiosity (1990) and also closely related to Predictability Minimization (1991)
基盤モデルの実ロボット応用に関するサーベイ論文を公開しました!
Meta AI ResearchのChrisさん
@chris_j_paxton
, Google DeepmindのAndyさん
@andyzeng_
という,この分野で最先端を進むお二人からのフィードバックを受けながら執筆できて大変貴重な経験になりました!
Can we solve Gym tasks with only 5-10 "trials"? Yes we can. We propose deployment efficiency as a new metric for RL, counting # of distinct data collection policies in learning. Most prior methods use 100s-1Ms. Joint w/
@Matsuo_Lab
@ofirnachum
1/
Our new work is open!!
We developed a "never-give-up" foundation-model-based service robot system.
It has a self-recovery mechanism by prompt updates though reasoning by LLMs and HRI.
Please watch the video how it works!!