Google DeepMind’s Post

Google DeepMind Is there any AI that can answer the following questions of market intelligence? "Who, in the Ontario province of Canada, have new US patents granted on the nearest Tuesday (Eastern Time), when the USPTO releases the newly granted US patents on a weekly basis?" "Who, in the "江蘇" province of China, have new US patents granted on the nearest Tuesday (Eastern Time), when the USPTO releases the newly granted US patents on a weekly basis?" With our intellectual property (IP), a Chinese-English multilingual metadata, we can get the full list answers for the above questions. Do you or any of your contacts need our expertise/IP to do the data analysis that AI can't do? Metadata is an enabler. It is like a treasure map for treasure hunting. Without metadata, like a treasure map, NO data can be found/retrieved, even by the most advanced technologies, like AI, high-end chips, supercomputers, etc. https://lnkd.in/g-aJFnXR

Embracing failure is essential to achieving success.

Well, like humans, sadly, robots fail as well. That's why it is important to explore robotic explanations for failures. Our paper about these explanations is published and available at IEEE now: https://lnkd.in/dCWkWQXH

These are ALOHA's behind-the-scenes videos! Thanks to the authors for an honest compilation of failure cases, as well as showing the full human remote control stack. Agile hardware platform is a critical step towards generalist robots. Let’s crack the AI challenge one algorithm at a time!

Love the honesty: behind the scene video of the amazing ALOHA robot. 🧑🔧🤖 This is the often undershared, bumpy progress towards an eventual sleek, shiny product that people absolutely love, one day. 🙌 "One algo at a time" it is!!! 🧑💻🧑🔧 See the "proper" video in the comment section. ⤵️ #aloha #ai #robot

[ICLR 2024] Our novel method DORAEMON brings zero-shot Sim-to-Real transfer one step closer! If you ever used Domain Randomization (DR) for transferring RL policies to real robots, you know how annoying it can be to manually tune these uniform distributions. With "Domain Randomization via Entropy Maximization", you only need to specify a desired performance threshold and it'll automatically find the widest DR distribution which still preserves such performance on the trained policy. Note: - No data / feedback from the real world is needed - We test in settings with up 17 randomized parameters - We test our method on 6 sim-to-sim benchmark tasks and 1 sim-to-real pushing task on the Franka Panda robot. Special thanks to Pascal Klink, Carlo D'Eramo, Georgia Chalvatzaki, Tatiana Tommasi, and Jan Peters for making this happen. Paper: https://lnkd.in/dmTtmezc Website: https://lnkd.in/dPzfuJqX Code: https://lnkd.in/dgVkje5M #reinforcementlearning #simtoreal #domainrandomization

A humanoid robot that can perform actions based on text prompts could pave the way for machines that behave more like us and communicate using gestures. Read more: https://lnkd.in/eYGR5Zi3

Figure Status Update! Figure 01 Humanoid Robot powered by OpenAl can now do full conversations with humans and perform tasks simultaneously

I created a new illustration for Kate Darling's tech articles, this one concerning the potential real-life reflections of violent behavior towards robots and chatbots. Client: BBC Science Focus AD: Joe Eden https://lnkd.in/dUYV_2R3 https://lnkd.in/dzvNVX4R #illustration #illustrationoftheday #editorialillustration #press #conceptualillustration #illo #chatbots #technology  #artificialintelligence #techillustration

Discover the Future of Robotics: Human-Mimicking Avatars 🚀🤖 Explore the groundbreaking world of advanced robots designed to mimic human movements and interactions! Delve into the cutting-edge technology behind these avatars, powered by sensors and artificial intelligence. Learn about the different types of sensors that enable precise movements and the role of AI and machine learning in teaching robots to perform tasks autonomously. Meet Beomni, Telesar, and NimbRo, the innovative avatars revolutionizing fields from healthcare and caregiving to space exploration and security. Witness how these robots enhance safety, efficiency, and companionship, paving the way for a future where human-robot collaboration becomes integral to daily life. #Robots #Robotics #HumanMimicking #ReGiuS #ReGiuSonYoutube #AI #Avatar #Avatars https://lnkd.in/ddQJV-qu