Loading video...

Video Failed to Load

Go Home

😈 Can we find your robot policy’s weaknesses before running it on hardware with different conditions (lighting, visual backdrop, distractors, etc.)? Thrilled to share my main sabbatical project Google DeepMind! Predictive Red Teaming: Breaking Policies Without Breaking Robots

Majumdar_Ani's profile picture

Anirudha Majumdar

7,675 subscribers

30,948 views • 1 year ago •via X (Twitter)

Science & TechnologyEducation
LIVE

Anya Rossi• Live Now

Private livecam show

0 Comments

No comments available

Comments from the original post will appear here

Related Videos

Robots are now so precise they can perform surgery on an egg without breaking It
0:17

Robots are now so precise they can perform surgery on an egg without breaking It

Dhakshana Moorthy

23,141 views • 9 months ago

Robots have become so accurate that they can carry out surgery on an egg without breaking it.
0:17

Robots have become so accurate that they can carry out surgery on an egg without breaking it.

Massimo

8,287,977 views • 4 months ago

How can we more effectively leverage robot data from different embodiments for skill transfer? Excited to share that our new work, RoVi-Aug, has been accepted to Conference on Robot Learning as an oral paper！ WIth RoVi-Aug, you can augment an existing robot dataset into a different robot and different viewpoints. A policy trained on the augmented dataset can zero-shot deploy on the unseen target robot with significantly different camera angles! 🧵👇 🔗 Check out our paper:
0:18

How can we more effectively leverage robot data from different embodiments for skill transfer? Excited to share that our new work, RoVi-Aug, has been accepted to Conference on Robot Learning as an oral paper！ WIth RoVi-Aug, you can augment an existing robot dataset into a different robot and different viewpoints. A policy trained on the augmented dataset can zero-shot deploy on the unseen target robot with significantly different camera angles! 🧵👇 🔗 Check out our paper:

Chenfeng_X

28,288 views • 1 year ago

Please stop breaking your spaghetti, there is a way to do it without breaking it..
0:10

Please stop breaking your spaghetti, there is a way to do it without breaking it..

Mickamious

274,037 views • 7 months ago

One more update 🎉 I’m so excited to share that I’ve joined Google DeepMind to lead design on AI Studio. The way we first interact with DeepMind’s incredible models needs to be a magical experience and I’m thrilled to shape that alongside this amazing team. Let’s go 🚀
0:17

One more update 🎉 I’m so excited to share that I’ve joined Google DeepMind to lead design on AI Studio. The way we first interact with DeepMind’s incredible models needs to be a magical experience and I’m thrilled to shape that alongside this amazing team. Let’s go 🚀

Ammaar Reshi

124,405 views • 1 year ago

Trying to work on my weakness. With time your weaknesses can become your strengths, best feeling ever !
0:12

Trying to work on my weakness. With time your weaknesses can become your strengths, best feeling ever !

Jacob “Pimp” Winneche

233,430 views • 10 months ago

Qualia has been selected for the Google DeepMind Robotics Program. We train embodied models that put a robot on a real manual task and make it work, on the floor, not in a demo. Foundation models and reasoning are where robotics is heading, and doing that work alongside DeepMind, who are pushing this frontier, is exactly where we want to be. If you are a company looking to see how a new generation of robots can help your manual tasks, contact us at hello@qualiastudios.dev More soon
0:16

Qualia has been selected for the Google DeepMind Robotics Program. We train embodied models that put a robot on a real manual task and make it work, on the floor, not in a demo. Foundation models and reasoning are where robotics is heading, and doing that work alongside DeepMind, who are pushing this frontier, is exactly where we want to be. If you are a company looking to see how a new generation of robots can help your manual tasks, contact us at [email protected] More soon

Qualia

87,429 views • 29 days ago

🎉 Happy to share my internship project Google DeepMind 🗼 – purely text-trained LLMs can act as evolutionary recombination operators 🦎 🧬 Our EvoLLM uses LLM backends to outperform competitive baselines. Work done w. Yingtao Tian & Yujin Tang 🤗 📜:
0:15

🎉 Happy to share my internship project Google DeepMind 🗼 – purely text-trained LLMs can act as evolutionary recombination operators 🦎 🧬 Our EvoLLM uses LLM backends to outperform competitive baselines. Work done w. Yingtao Tian & Yujin Tang 🤗 📜:

Robert Lange

39,076 views • 2 years ago

had fun working on my first hardware project recently with my son, we built a little radio transmitter that sends messages into space! we hooked it up to if you want to send one
0:14

had fun working on my first hardware project recently with my son, we built a little radio transmitter that sends messages into space! we hooked it up to if you want to send one

saint laurent del rey

34,916 views • 4 months ago

RoboBallet: End of Manual Robot Programming Google DeepMind/Intrinsic/UCL just cracked multi-robot coordination. Published in Science Robotics. 🤖What it does: – 8 robots, 40 tasks, zero collisions - all planned in seconds – 25% better motion plans than expert-designed solutions – 60% faster task completion when scaling from 4 to 8 robots – Works on factory layouts, it's never seen 👩‍💻 The tech: – Graph Neural Networks map robots/obstacles as connected nodes – Reinforcement learning trains on millions of scenarios – No manual programming required - just give it CAD files and tasks ❓Question: The bottleneck to robot collaboration has always been manual prototyping. Now that we can automate the automation what will this unlock?
0:17

RoboBallet: End of Manual Robot Programming Google DeepMind/Intrinsic/UCL just cracked multi-robot coordination. Published in Science Robotics. 🤖What it does: – 8 robots, 40 tasks, zero collisions - all planned in seconds – 25% better motion plans than expert-designed solutions – 60% faster task completion when scaling from 4 to 8 robots – Works on factory layouts, it's never seen 👩‍💻 The tech: – Graph Neural Networks map robots/obstacles as connected nodes – Reinforcement learning trains on millions of scenarios – No manual programming required - just give it CAD files and tasks ❓Question: The bottleneck to robot collaboration has always been manual prototyping. Now that we can automate the automation what will this unlock?

Jack 🤖

41,633 views • 10 months ago

🎉 Congratulations Google DeepMind on the launch of Gemma 3 🎊 Introducing lightweight, multimodal, multilingual models streamlined for optimal performance on all your NVIDIA hardware – from your #datacenter GPUs to Jetson and RTX AI PCs for Windows. ➡️
0:18

🎉 Congratulations Google DeepMind on the launch of Gemma 3 🎊 Introducing lightweight, multimodal, multilingual models streamlined for optimal performance on all your NVIDIA hardware – from your #datacenter GPUs to Jetson and RTX AI PCs for Windows. ➡️

NVIDIA AI Developer

13,828 views • 1 year ago

Force feedback demo Force feedback is when joystick is pushing on your hand when something is pushing on the robot arm. Feeling the force - so much helpful to control the robot, that done well it allows you to do tasks even without visual feed. You can make an experiment: close your eyes - you can easily get the headphones out of the case. Also, visual information is often not enough. For example, you're trying to pull out a usb connector, but you pull it at the wrong angle, causing it to get stuck. Visually, nothing changes, but the pressure is intense and you can break the connector. Surgical robots have been using force feedback for years, and there are also 3D styluses which use this feature, proving that the technology works and is useful. But in modern robots with AI, it's hardly ever implemented. Although it's useful for both teleoperation and AI model. That's one of the reasons why we are building our robotic arms starting with off the shelf motors rather than taking the whole off the shelf arm. There are still a range of easy wins that can be made iterating robot hardware.
0:15

Force feedback demo Force feedback is when joystick is pushing on your hand when something is pushing on the robot arm. Feeling the force - so much helpful to control the robot, that done well it allows you to do tasks even without visual feed. You can make an experiment: close your eyes - you can easily get the headphones out of the case. Also, visual information is often not enough. For example, you're trying to pull out a usb connector, but you pull it at the wrong angle, causing it to get stuck. Visually, nothing changes, but the pressure is intense and you can break the connector. Surgical robots have been using force feedback for years, and there are also 3D styluses which use this feature, proving that the technology works and is useful. But in modern robots with AI, it's hardly ever implemented. Although it's useful for both teleoperation and AI model. That's one of the reasons why we are building our robotic arms starting with off the shelf motors rather than taking the whole off the shelf arm. There are still a range of easy wins that can be made iterating robot hardware.

Igor Kulakov

18,773 views • 1 year ago

Evaluating policies on a real robot can be painful. Can we use a world model to get a rough estimate of how good a policy is? Checkout "Evaluating Robot Policies in a World Model". Paper: Demo: Code:
0:10

Evaluating policies on a real robot can be painful. Can we use a world model to get a rough estimate of how good a policy is? Checkout "Evaluating Robot Policies in a World Model". Paper: Demo: Code:

Sherry Yang

36,568 views • 1 year ago

BREAKING 🚨: Google is working on "Flashcards" for NotebookLM and a possibility to discover sources across your Google Drive! "Generate AI flashcards based on your sources"
0:19

BREAKING 🚨: Google is working on "Flashcards" for NotebookLM and a possibility to discover sources across your Google Drive! "Generate AI flashcards based on your sources"

TestingCatalog News 🗞

47,842 views • 1 year ago

🚨 BREAKING! The first ever X comedy thread is here! Different genres of comedy and things to bring ten your day! Enjoy and share with others!
0:11

🚨 BREAKING! The first ever X comedy thread is here! Different genres of comedy and things to bring ten your day! Enjoy and share with others!

Liberal Tear Creator™

15,472 views • 7 months ago

RobustNeRF: Ignoring Distractors with Robust Losses abs: project page:
0:18

RobustNeRF: Ignoring Distractors with Robust Losses abs: project page:

AK

65,606 views • 3 years ago

🔥🚨BREAKING: Women will be having more sex with ROBOTS than men before the end of this year according to “The Sun.”
0:17

Sensitive content

🔥🚨BREAKING: Women will be having more sex with ROBOTS than men before the end of this year according to “The Sun.”

Dom Lucre | Breaker of Narratives

313,741 views • 1 year ago

Seedance 2.0 is also quite strong on some visual reasoning tasks — I just told it to place the geometric boards in the right positions. Of courses, it is not perfect l. Image from Google deepmind.
0:12

Seedance 2.0 is also quite strong on some visual reasoning tasks — I just told it to place the geometric boards in the right positions. Of courses, it is not perfect l. Image from Google deepmind.

padphone

21,763 views • 4 months ago

Wouldn’t it be nice if ChatGPT could find your missing keys for you? Our latest research from Berkeley AI Research + Google AI suggests that robots can use large language models (LLMs) to find hidden objects faster. 🧵👇
0:15

Wouldn’t it be nice if ChatGPT could find your missing keys for you? Our latest research from Berkeley AI Research + Google AI suggests that robots can use large language models (LLMs) to find hidden objects faster. 🧵👇

Kaushik Shivakumar

58,508 views • 3 years ago

BREAKING: Google just turned Gemini into a full-blown AI school system. Teachers can now assign AI experts to students. Students can auto-generate quizzes and visual explainers. And it's all free in Google Workspace for Education. Here’s what just dropped 👇
0:14

BREAKING: Google just turned Gemini into a full-blown AI school system. Teachers can now assign AI experts to students. Students can auto-generate quizzes and visual explainers. And it's all free in Google Workspace for Education. Here’s what just dropped 👇

Brendan Jowett

847,704 views • 1 year ago