正在加载视频...

视频加载失败

加载此视频时出现问题。这可能是由于临时网络问题，或视频可能不可用。

One less-known fact about glove-based data collection: it produces higher quality data than teleop on contact-rich tasks. Remote teleop can’t provide good force feedback, but gloves do naturally, making tasks like sock folding, which rely on feel, far easier to capture.

Tony Zhao

85,556 subscribers

185,996 次观看 • 7 个月前 •via X (Twitter)

教育科学技术

Anya Rossi• Live Now

Private livecam show

0 条评论

暂无评论

原始帖子的评论将显示在这里

相关视频

Low-cost teleop systems have democratized robot data collection, but they lack any force feedback, making it challenging to teleoperate contact-rich tasks. Many robot arms provide force information — a critical yet underutilized modality in robot learning. We introduce: 1. 🦾A low-cost, force-feedback-enabled teleop system. 2. 🥊Force-Attending Curriculum Training (FACTR) uses force to improve generalization in complex, contact-rich tasks. 🧵(1/N)

Low-cost teleop systems have democratized robot data collection, but they lack any force feedback, making it challenging to teleoperate contact-rich tasks. Many robot arms provide force information — a critical yet underutilized modality in robot learning. We introduce: 1. 🦾A low-cost, force-feedback-enabled teleop system. 2. 🥊Force-Attending Curriculum Training (FACTR) uses force to improve generalization in complex, contact-rich tasks. 🧵(1/N)

Jason Liu

151,595 次观看 • 1 年前

So I heard we need more data for robot learning :) Purely real world teleop is expensive and slow, making large scale data collection challenging. I’ve been excited about getting more data into robot learning, going beyond just real-world teleop data. To this end, we’ve been scaling up data generation with RL in realistic simulations generated on the fly from crowdsourced videos. Enables realistic data collection, much more cheaply than purely real world teleop. Importantly, data collection becomes even*cheaper* with more environments, allowing training with over 100x more data. Transfers to real robots for generalizable manipulation. A 🧵 (1/N)

So I heard we need more data for robot learning :) Purely real world teleop is expensive and slow, making large scale data collection challenging. I’ve been excited about getting more data into robot learning, going beyond just real-world teleop data. To this end, we’ve been scaling up data generation with RL in realistic simulations generated on the fly from crowdsourced videos. Enables realistic data collection, much more cheaply than purely real world teleop. Importantly, data collection becomes evencheaper with more environments, allowing training with over 100x more data. Transfers to real robots for generalizable manipulation. A 🧵 (1/N)

Abhishek Gupta

13,346 次观看 • 1 年前

𝗢𝗽𝗲𝗻𝗔𝗿𝗺 is an open-source robot arm designed for human manipulation data collection. What's new in 𝘃𝟬.𝟮 (beta): ・Teleop with gravity compensation ・Bilateral force feedback support #OpenSource #Robotics #Humanoids #ArtificialIntelligence

𝗢𝗽𝗲𝗻𝗔𝗿𝗺 is an open-source robot arm designed for human manipulation data collection. What's new in 𝘃𝟬.𝟮 (beta): ・Teleop with gravity compensation ・Bilateral force feedback support #OpenSource #Robotics #Humanoids #ArtificialIntelligence

Reazon Human Interaction Lab

43,128 次观看 • 1 年前

On-chain automation just leveled up. We’ve added data hooks to Orbit, making it easier than ever to set up trades or automate on-chain tasks. Here’s why this changes the game

On-chain automation just leveled up. We’ve added data hooks to Orbit, making it easier than ever to set up trades or automate on-chain tasks. Here’s why this changes the game

Orbit

39,858 次观看 • 1 年前

Thanks Yunfang fr Sharpa walking me thro these incredibly dexterous robotic hands. This particular demo is teleop; but still really impressive. Evidently the teleoperator do get force feedback on his teleop device (very much needed to card dealing) #IROS2025

Thanks Yunfang fr Sharpa walking me thro these incredibly dexterous robotic hands. This particular demo is teleop; but still really impressive. Evidently the teleoperator do get force feedback on his teleop device (very much needed to card dealing) #IROS2025

Michael Cho - Rbt/Acc

25,744 次观看 • 8 个月前

Thoughts from a teleop session today: 1/ Teleop is painful. UMI-style grippers are making more and more sense to me: shorter per-episode execution, more data, and more intuitive for factory workers who will be the ones with my product eventually. Wondering if PI resists this because researchers aren't the ones collecting the data? 2/ The take-away of the HF shirt folding post stuck with me: data quality matters most of all. I'm 30 minutes into this task and still making mistakes with teleop. What’s the perspective for non-roboticists? Maybe VR headset is better. Want to try that next. 3/ I’m noticeable better at teleop even when I’m just 15cm closer. 4/ Double-close gestures for re-record (left) and early episode end (right) are a game changer. Credit Andreas Köpf. 5/ Want to gamify my own collection more: thinking of a daily target dashboard. 6/ I’d like to rate each episode with a 1-5 data quality score. Don't wanna throw away bad data away, but still be able to filter top-quality. Maybe possible with foot pedals?

Thoughts from a teleop session today: 1/ Teleop is painful. UMI-style grippers are making more and more sense to me: shorter per-episode execution, more data, and more intuitive for factory workers who will be the ones with my product eventually. Wondering if PI resists this because researchers aren't the ones collecting the data? 2/ The take-away of the HF shirt folding post stuck with me: data quality matters most of all. I'm 30 minutes into this task and still making mistakes with teleop. What’s the perspective for non-roboticists? Maybe VR headset is better. Want to try that next. 3/ I’m noticeable better at teleop even when I’m just 15cm closer. 4/ Double-close gestures for re-record (left) and early episode end (right) are a game changer. Credit Andreas Köpf. 5/ Want to gamify my own collection more: thinking of a daily target dashboard. 6/ I’d like to rate each episode with a 1-5 data quality score. Don't wanna throw away bad data away, but still be able to filter top-quality. Maybe possible with foot pedals?

Dominique Paul

37,908 次观看 • 2 个月前

Nice one! We also worked on cloth folding sim2real last year with a custom solver for massive parallelization. The results might look less decent but it is teleop-free. Release soon!

Nice one! We also worked on cloth folding sim2real last year with a custom solver for massive parallelization. The results might look less decent but it is teleop-free. Release soon!

C. Zhang

17,915 次观看 • 2 个月前

OpenArm is a fully open-source robotic arm designed for human manipulation data collection - developed by the Reazon Human Interaction Lab in Japan. V0.2 beta features smoother teleop with gravity compensation and bilateral force feedback.

OpenArm is a fully open-source robotic arm designed for human manipulation data collection - developed by the Reazon Human Interaction Lab in Japan. V0.2 beta features smoother teleop with gravity compensation and bilateral force feedback.

The Humanoid Hub

38,190 次观看 • 1 年前

Force is arguably the most overlooked ingredient in modern robot learning. Introducing FACTR 2: it turns *any* commodity robot into a force-aware system with no force sensors required. Train a tiny force network in <1min with <10mins of data and drop it into any existing teleop pipelines: ✅ Free force sensing for both the robot and the operator arm ✅ Makes demos higher-quality → fewer of them needed. ✅ A new force-aware learning algorithm (FIRST) uses those recovered forces to figure out which parts of a demo actually matter, making learning data-efficient. ✅ Strong performance on complex tasks with fewer demos and even no pretraining! More details below.

Force is arguably the most overlooked ingredient in modern robot learning. Introducing FACTR 2: it turns any commodity robot into a force-aware system with no force sensors required. Train a tiny force network in <1min with <10mins of data and drop it into any existing teleop pipelines: ✅ Free force sensing for both the robot and the operator arm ✅ Makes demos higher-quality → fewer of them needed. ✅ A new force-aware learning algorithm (FIRST) uses those recovered forces to figure out which parts of a demo actually matter, making learning data-efficient. ✅ Strong performance on complex tasks with fewer demos and even no pretraining! More details below.

Deepak Pathak

38,178 次观看 • 23 天前

Most glove VR devices suffer from the problem of gloves having a significant negative impact on your finger dexterity. This is one of the coolest things about Martino Russi's hand glove design - for teleop, data collection, or even VR. Your fingertips remain uncovered and let you fully manipulate small objects with precision - save the clunkiness of wearing the device on your hand. Thinking of some comfort and durability upgrades here.. Lots of experimentation we gotta do.

Most glove VR devices suffer from the problem of gloves having a significant negative impact on your finger dexterity. This is one of the coolest things about Martino Russi's hand glove design - for teleop, data collection, or even VR. Your fingertips remain uncovered and let you fully manipulate small objects with precision - save the clunkiness of wearing the device on your hand. Thinking of some comfort and durability upgrades here.. Lots of experimentation we gotta do.

Sam Padilla

25,317 次观看 • 1 年前

Impressive on every dimension! Genesis AI has announced GENE-26.5, its first robotic foundation model, aimed at achieving human-level dexterous manipulation: - Single model, shared weights, handles egg cracking, lab pipetting, bimanual Rubik's Cube, smoothie making, and wire harnessing. Most tasks: under 1 hour of task-specific data, 1x real-world speed. - Genesis Hand 1.0: human-size, 20 active DoF, soft-contact skin. Paired glove gives a 1:1:1 mapping (glove, human hand, robot hand). 100x cheaper than teleop, 5x more data-efficient internally. - Data engine pulls from three sources: glove data, egocentric (head-cam) video, and large-scale internet video. The 1:1 hand-to-human match closes the embodiment gap, letting Genesis use video data more effectively than rivals. - Flow matching across vision, tactile, proprioception, and language inputs. Per Genesis, scaling data and compute directly improves zero-shot

Impressive on every dimension! Genesis AI has announced GENE-26.5, its first robotic foundation model, aimed at achieving human-level dexterous manipulation: - Single model, shared weights, handles egg cracking, lab pipetting, bimanual Rubik's Cube, smoothie making, and wire harnessing. Most tasks: under 1 hour of task-specific data, 1x real-world speed. - Genesis Hand 1.0: human-size, 20 active DoF, soft-contact skin. Paired glove gives a 1:1:1 mapping (glove, human hand, robot hand). 100x cheaper than teleop, 5x more data-efficient internally. - Data engine pulls from three sources: glove data, egocentric (head-cam) video, and large-scale internet video. The 1:1 hand-to-human match closes the embodiment gap, letting Genesis use video data more effectively than rivals. - Flow matching across vision, tactile, proprioception, and language inputs. Per Genesis, scaling data and compute directly improves zero-shot

The Humanoid Hub

95,119 次观看 • 1 个月前

The most frustrating part of imitation learning is collecting huge amounts of teleop data. But why teleop robots when robots can learn by watching us? Introducing Point Policy, a novel framework that enables robots to learn from human videos without any teleop, sim2real, or RL.

The most frustrating part of imitation learning is collecting huge amounts of teleop data. But why teleop robots when robots can learn by watching us? Introducing Point Policy, a novel framework that enables robots to learn from human videos without any teleop, sim2real, or RL.

Siddhant Haldar

69,056 次观看 • 1 年前

ManiFeel is a benchmark for contact-rich manipulation tasks. Force- and tactile-based tasks with occlusions and partially-observable state are going to be crucial for true general purpose robotics. Website:

ManiFeel is a benchmark for contact-rich manipulation tasks. Force- and tactile-based tasks with occlusions and partially-observable state are going to be crucial for true general purpose robotics. Website:

Chris Paxton

12,063 次观看 • 4 个月前

1/ Create a spreadsheet, analyze the data in Excel, and then schedule tasks based on the output

1/ Create a spreadsheet, analyze the data in Excel, and then schedule tasks based on the output

Satya Nadella

162,558 次观看 • 4 个月前

🚨🚨New paper: if you want robots to do bimanual long-horizon tasks well, try RaC: a human in-the-loop data collection protocol that naturally amplifies Recovery & Correction behaviors + trains on them. 📈 data efficiency 10x vs prior results (+many nice properties). 🧵⬇️

🚨🚨New paper: if you want robots to do bimanual long-horizon tasks well, try RaC: a human in-the-loop data collection protocol that naturally amplifies Recovery & Correction behaviors + trains on them. 📈 data efficiency 10x vs prior results (+many nice properties). 🧵⬇️

Aviral Kumar

25,884 次观看 • 9 个月前

NASA's repository of geospatial data contains key insights on everything from climate and air quality to urban planning and disaster response—but its scale and complexity make it difficult to parse. With Earth Copilot, they're making it easier than ever for anyone to access that data:

NASA's repository of geospatial data contains key insights on everything from climate and air quality to urban planning and disaster response—but its scale and complexity make it difficult to parse. With Earth Copilot, they're making it easier than ever for anyone to access that data:

Satya Nadella

107,972 次观看 • 1 年前

Humanoid robotics is hitting a data wall. Teleop and mocap took us far, but they don’t scale to every object, terrain, and behavior. We’re releasing GRAIL: — a fully digital pipeline for generating loco-manipulation data before the robot moves. 🧵(1/8)

Humanoid robotics is hitting a data wall. Teleop and mocap took us far, but they don’t scale to every object, terrain, and behavior. We’re releasing GRAIL: — a fully digital pipeline for generating loco-manipulation data before the robot moves. 🧵(1/8)

Ye Yuan

42,615 次观看 • 1 个月前

Our newest model, π0.7, has some interesting emergent capabilities: it can control a new robot to fold shirts for which we had no shirt folding data, figure out how to use an appliance with language-based coaching, and perform a wide range of dexterous tasks all in one model!

Our newest model, π0.7, has some interesting emergent capabilities: it can control a new robot to fold shirts for which we had no shirt folding data, figure out how to use an appliance with language-based coaching, and perform a wide range of dexterous tasks all in one model!

Physical Intelligence

457,802 次观看 • 2 个月前

Lumos data collection tools -- quality tools and software like this will make it possible to scale quality data way more easily

Lumos data collection tools -- quality tools and software like this will make it possible to scale quality data way more easily

Chris Paxton

15,747 次观看 • 5 个月前