Google DeepMind has unveiled its latest project, ALOHA Unleashed, which demonstrates significant advancements in robotic dexterity and autonomy. This project, developed over the past year, features low-cost robots capable of performing highly complex tasks such as hanging shirts on hangers, robot repair, and tying shoelaces. These robots, part of the ALOHA 2 fleet, operate autonomously and have been showcased in videos that highlight their ability to execute tasks in a single continuous shot, proving the robustness of their policies. The robots' persistence and capability to adjust their actions to complete tasks have been emphasized in recent demonstrations. Additionally, an open-source AI initiative involving DJI robots controlled by models like Whisper, Idefics2, and Parler-TTS has been shared, complete with a 250 lines code gist for community development.
Excited to share one of the last projects I worked on at Google -- We picked three of the most dextrous tasks we could think of: - tying shoelaces 👟 - replacing a robot finger 🤖 - hanging a shirt 👕 and tried to see if we could train a policy to do each. All of them worked! https://t.co/pyinM151Tj
Google DeepMind's autonomous ALOHA robots can perform robot repair, tie shoes and hang shirts https://t.co/Khfl3qbS95
really incredible progress on highly dextrous manipulation tasks, all shown in a single take to show how robust the policies are. Congrats @tonyzzhao, @ayzwah and team! https://t.co/PViCz32EFg