That's very cool! A little over a year ago, I started on a large bot(currently weighing in at about 16 pounds) with the intent of turning it over to something akin to Claude. You did a fantastic job coming up with this.
I’m actually very interested in imitation learning and am building a system to do exactly this. It can be done much cheaper now, probably for about $7k total. I’ve replicated ALOHA’s ACT results using a cheap $250 table top arm but am very interested in doing something bigger, more complex, and useful. RoBart originally started as a way to build a cheap mobile base for a robot that could be operated and trained at home. Turns out that the hoverboard based set up isn’t quite optimal so for fun, I wanted to see how far I could push GPT-4 and Claude to autonomously operate it.
Did you find it useful? Did you use any other ML at the same time
I think about the different aims of of YOLO object recognition vs LLM conversational vs RL for planning and how one might integrate them for a better overall system
Re: LLM for supervising an imitation learning policy, check out the recent paper “Robot Utility Models” :) They even bought a TLD for it: https://robotutilitymodels.com/
9 comments
[ 0.21 ms ] story [ 49.2 ms ] threadHave you seen the Aloha robot? Would be cool to see LLMs added to that, the price tag is a bit more at ~$35k...
https://mobile-aloha.github.io/
I think about the different aims of of YOLO object recognition vs LLM conversational vs RL for planning and how one might integrate them for a better overall system
Re: LLM for supervising an imitation learning policy, check out the recent paper “Robot Utility Models” :) They even bought a TLD for it: https://robotutilitymodels.com/