Younghyo Park's banner

Younghyo Park

@younghyo_park • 2,101 subscribers

Robotics Ph.D. student at @MIT_CSAIL, advised by @pulkitology

Shorts

[8/n] By the way, for the simulation folks wondering how we simulated tight-clearance parts: the usual tools didn't work for us. SDF was so slow and hard to tune. Off-the-shelf convex decomposition like V-HACD or CoACD filled the clearance, so parts that should fit jam. So we built mjfit. It decomposes common assembly parts (pegs and holes of any outline, threads, and gears) into convex pieces that keep the clearance exactly as designed, and it runs an order of magnitude faster than SDF. It's open source:

[8/n] By the way, for the simulation folks wondering how we simulated tight-clearance parts: the usual tools didn't work for us. SDF was so slow and hard to tune. Off-the-shelf convex decomposition like V-HACD or CoACD filled the clearance, so parts that should fit jam. So we built mjfit. It decomposes common assembly parts (pegs and holes of any outline, threads, and gears) into convex pieces that keep the clearance exactly as designed, and it runs an order of magnitude faster than SDF. It's open source:

28,869 görüntüleme

What's different between these two BC policies? It's the same architecture, training budget, and data collection setup — the only difference is the controller gains! Controller gains are an understudied design parameter in robot learning. In our new work (w/ Antonia Bronars*, Pulkit Agrawal), we show how they act as an inductive bias across BC, RL, and Sim2Real transfer, with real consequences on performance. Here's what we found 🧵 * Equal Contribution 📄arxiv: 🔗website:

What's different between these two BC policies? It's the same architecture, training budget, and data collection setup — the only difference is the controller gains! Controller gains are an understudied design parameter in robot learning. In our new work (w/ Antonia Bronars, Pulkit Agrawal), we show how they act as an inductive bias across BC, RL, and Sim2Real transfer, with real consequences on performance. Here's what we found 🧵 Equal Contribution 📄arxiv: 🔗website:

164,822 görüntüleme

Nobody asked but I'm going to talk about that cuboid in the video anyway :) It's a 3D-printed, 6DoF-pose-trackable cuboid, programmatically generated and tracked with this! This little cube really helped us isolate control issues from RGB vision problems and clearly realize the importance of low-level controller design decisions.

Nobody asked but I'm going to talk about that cuboid in the video anyway :) It's a 3D-printed, 6DoF-pose-trackable cuboid, programmatically generated and tracked with this! This little cube really helped us isolate control issues from RGB vision problems and clearly realize the importance of low-level controller design decisions.

81,905 görüntüleme

🥽 Want to use your new Apple Vision Pro to control your robot? Want to record how you navigate / manipulate the world to train a policy? I developed an app for VisionOS that can stream your head / wrist / finger movements over WiFi, which you can subscribe on any machines using a simple python library. 🧵[1/6]

🥽 Want to use your new Apple Vision Pro to control your robot? Want to record how you navigate / manipulate the world to train a policy? I developed an app for VisionOS that can stream your head / wrist / finger movements over WiFi, which you can subscribe on any machines using a simple python library. 🧵[1/6]

214,278 görüntüleme

Videos

Anya Rossi

sweetdream.ai

SweetDream.ai•Sponsored•Livecam

Watch Anya Live

Anya is streaming live right now! Join her private show and enjoy exclusive content.

Exclusive private shows

1.2k viewers online

Private Show

Join now for exclusive access

Free preview available • Premium content

Can a robot assemble tight-clearance parts using only its proprioception? No camera, no tactile sensor, no F/T sensor, no joint torque sensing, no priors over the object geometries or poses: just joint position readings. It actually can! Blind policies trained in simulation insert pegs, mesh gears, and thread nuts, and the peg policy works zero-shot in the real world. Here's what we found 🧵 Joint work with Antonia Bronars, my wonderful co-first author, who's on the job market right now! Advised by Kushal and Pulkit Agrawal.

Can a robot assemble tight-clearance parts using only its proprioception? No camera, no tactile sensor, no F/T sensor, no joint torque sensing, no priors over the object geometries or poses: just joint position readings. It actually can! Blind policies trained in simulation insert pegs, mesh gears, and thread nuts, and the peg policy works zero-shot in the real world. Here's what we found 🧵 Joint work with Antonia Bronars, my wonderful co-first author, who's on the job market right now! Advised by Kushal and Pulkit Agrawal.

36,261 görüntüleme • 2 gün önce

Collect robot demos from anywhere through AR! Excited to introduce 🎯DART, Dexterous AR Teleoperation interface enabling anyone to teleoperate robots in cloud-hosted simulation. With DART, anyone can collect robot demos anywhere, anytime, for multiple robots and tasks in one sitting. Every data is automatically logged on our open-sourced cloud database DexHub for public use. 🧵[1/n]

Collect robot demos from anywhere through AR! Excited to introduce 🎯DART, Dexterous AR Teleoperation interface enabling anyone to teleoperate robots in cloud-hosted simulation. With DART, anyone can collect robot demos anywhere, anytime, for multiple robots and tasks in one sitting. Every data is automatically logged on our open-sourced cloud database DexHub for public use. 🧵[1/n]

27,630 görüntüleme • 1 yıl önce

Daha fazla içerik yok.