Video yükleniyor...

Video Yüklenemedi

Bu video yüklenirken bir sorun oluştu. Bu geçici bir ağ sorunundan kaynaklanıyor olabilir veya video kullanılamıyor olabilir.

Ana Sayfaya Dön

1/5 📢 Excited to share PowerSim, which lets us (a) interact with objects in captured scenes, (b) simulate how they move and deform, and (c) render their changing appearance. This is one of the first papers from the first cohort of PhD students in my new 3P Vision Group... show more

Anand Bhattad

3,701 subscribers

13,478 görüntüleme • 1 ay önce •via X (Twitter)

Bilim & Teknoloji Eğitim

Anya Rossi• Live Now

Private livecam show

6 Yorum

Anand Bhattad profil fotoğrafı

Anand Bhattad1 ay önce

2/5 Why did we build PowerSim? We want to capture scenes and then push, pull, or rearrange the objects in them. But what happens to their appearance when they deform? While working on another project, we ran into limitations of PhysGaussian and other existing methods under large deformations. You can see some of these in the video below: surface details disappear or become blurry as the object stretches. We needed a better way to handle this.

Anand Bhattad profil fotoğrafı

Anand Bhattad1 ay önce

3/5 What do we do? We build on Power Foam, whose primitives partition space into volumetric cells and combine them with Material Point Method (MPM) simulation. The connection is quite natural here. MPM tracks how an object deforms and we can use that information to update the primitives’ geometry and appearance without converting the scene into any other representation. A big shout-out to the Power Foam authors (@Shrisudhan001, Daniel Rebain, @kwangmoo_yi, @taiyasaki, @dorverbin, @prabhubility99) for the foundation this work builds on!

Anand Bhattad profil fotoğrafı

Anand Bhattad1 ay önce

4/5 Simply moving the PowerFoam primitives with MPM isn’t enough. Here, we twist a bread model, treating it as an elastic material. If we leave the dipole plane or detail axis fixed, the surface appearance doesn’t follow the deformation correctly. PowerSim rotates both with the material, preserving detail as it twists.

Anand Bhattad profil fotoğrafı

Anand Bhattad1 ay önce

5/5 With PowerSim, we can: - Physically interact with objects in captured scenes. - Recover material properties that vary across an object. - Combine objects from separately captured scenes and simulate them together. - Ray-trace reflections that change as objects move and deform. Huge congrats to @Tung11199951 for bringing this together! Excited to see what others build with PowerSim 😊

Anand Bhattad profil fotoğrafı

Anand Bhattad1 ay önce

6/5: PowerSim code is now live! Give it a try — we would love to see what you build with it. 💻

Lantern Works profil fotoğrafı

Lantern Works1 ay önce

My game Halls of History runs on splat rooms in the browser, and your 2/5 explains why PowerSim moved off splats. How close does it get to interactive rates? I couldn't find a runtime in the paper.

Benzer Videolar

📢 New paper alert 📢 Excited to share the first-author paper of Kunal Rai, PhD scholar at IITH in PNASNews We zoomed into the cellular gatekeeper and measured cooperative glycerol permeability for the first time using MD simulation #Biophysics

📢 New paper alert 📢 Excited to share the first-author paper of Kunal Rai, PhD scholar at IITH in PNASNews We zoomed into the cellular gatekeeper and measured cooperative glycerol permeability for the first time using MD simulation #Biophysics

Himanshu Joshi

11,942 görüntüleme • 2 ay önce

9PM first pitch. In the meantime enjoy this behind the scenes rain delay tour which ends with a surprise appearance by a broadcasting legend and one of my favorite people to see on the road.

9PM first pitch. In the meantime enjoy this behind the scenes rain delay tour which ends with a surprise appearance by a broadcasting legend and one of my favorite people to see on the road.

CJ

30,508 görüntüleme • 2 ay önce

PhD Students – How to generate the first draft of ANY type of literature review in 10 minutes? Here is how it works. 1. Go to 2. Enter the topic of your literature review 3. Set the filters e.g., type of papers and year, etc. 4. Get the first draft of your literature review 5. Work on it to refine and contextualize the draft

PhD Students – How to generate the first draft of ANY type of literature review in 10 minutes? Here is how it works. 1. Go to 2. Enter the topic of your literature review 3. Set the filters e.g., type of papers and year, etc. 4. Get the first draft of your literature review 5. Work on it to refine and contextualize the draft

Faheem Ullah

23,259 görüntüleme • 6 ay önce

This is FacePoke AI that lets you move the face and change expressions of portrait Wild demos (and how to try) 1. Girl with Pearl Earrings

This is FacePoke AI that lets you move the face and change expressions of portrait Wild demos (and how to try) 1. Girl with Pearl Earrings

Min Choi

149,867 görüntüleme • 2 yıl önce

📢📢📢 Excited to share our ICLR25 work, ArtGS, which can reconstruct articulated objects and scenes from human-scene interactions. 💻🖱️🧱🪑📖 Understanding physics and articulation is extremely important. With ArtGS, we can build a digital and simulatable replica of the real 3D world with minimal human effort!

📢📢📢 Excited to share our ICLR25 work, ArtGS, which can reconstruct articulated objects and scenes from human-scene interactions. 💻🖱️🧱🪑📖 Understanding physics and articulation is extremely important. With ArtGS, we can build a digital and simulatable replica of the real 3D world with minimal human effort!

Siyuan Huang

12,833 görüntüleme • 1 yıl önce

Isha's Scholarships for Tribal Students - September 29, 2024 Hope in their eyes, dreams in their hearts. For the tribal students near #IshaYogaCenter Coimbatore, a scholarship is more than just financial aid - it's the bridge to a world of opportunities they never thought possible. The first in their families to hold a book, sit in a classroom, and speak in English, they are breaking the chains of hardship and rewriting their destiny. Listen in as they share how this one opportunity changed their life. #IshaFoundation

Isha's Scholarships for Tribal Students - September 29, 2024 Hope in their eyes, dreams in their hearts. For the tribal students near #IshaYogaCenter Coimbatore, a scholarship is more than just financial aid - it's the bridge to a world of opportunities they never thought possible. The first in their families to hold a book, sit in a classroom, and speak in English, they are breaking the chains of hardship and rewriting their destiny. Listen in as they share how this one opportunity changed their life. #IshaFoundation

Isha Foundation

152,834 görüntüleme • 2 yıl önce

One of my favorite teaching tips to share: Test Talk. For the first 5 minutes of a testing session, students put their writing utensils on the ground, I hand out the tests, and students can collaborate. I hear some of the best mathematical conversations in these 5 minutes.

One of my favorite teaching tips to share: Test Talk. For the first 5 minutes of a testing session, students put their writing utensils on the ground, I hand out the tests, and students can collaborate. I hear some of the best mathematical conversations in these 5 minutes.

Howie Hua

51,500 görüntüleme • 2 yıl önce

The surprise is out WERE TAKING THE IRLS TO ITALY IN A FEW DAYS!!!! 🇮🇹 so excited to share our first time in Italy together and with chat there with us 🥹

The surprise is out WERE TAKING THE IRLS TO ITALY IN A FEW DAYS!!!! 🇮🇹 so excited to share our first time in Italy together and with chat there with us 🥹

Clara 🦋

57,065 görüntüleme • 1 yıl önce

NINE YEARS in the industry and this is the FIRST song they made as a group and they still move like it was released yesterday KIDS PLEASE 🥹🥹🥹🥹🥹

NINE YEARS in the industry and this is the FIRST song they made as a group and they still move like it was released yesterday KIDS PLEASE 🥹🥹🥹🥹🥹

BestFriend

113,220 görüntüleme • 3 ay önce

This week, I overhauled concurrent heap defragmentation (akin to garbage collection) in my voxel game, achieving microsecond-level pause times and O(1) scalability across all platforms. Since existing research didn't fit my needs, I built my own defragmentation algorithm from first principles. The new system is a game-changer for the Gamecube and Dreamcast, where memory can easily become a bottleneck. These improvements eliminate stutters, even with millions of allocated objects like chunks and entities. In the first clip, the PC version runs 100,000 objects (render distance: 32) with no issues. The second clip in the reply ramps up to half a million objects (render distance: 64) and still holds steady — pause times stayed rock solid.

This week, I overhauled concurrent heap defragmentation (akin to garbage collection) in my voxel game, achieving microsecond-level pause times and O(1) scalability across all platforms. Since existing research didn't fit my needs, I built my own defragmentation algorithm from first principles. The new system is a game-changer for the Gamecube and Dreamcast, where memory can easily become a bottleneck. These improvements eliminate stutters, even with millions of allocated objects like chunks and entities. In the first clip, the PC version runs 100,000 objects (render distance: 32) with no issues. The second clip in the reply ramps up to half a million objects (render distance: 64) and still holds steady — pause times stayed rock solid.

A Flock of Meese

38,548 görüntüleme • 1 yıl önce

One of my favorite scenes in The 13th Warrior is when the Arab ambassadors arrive at the funeral feast and try to communicate with the Northmen in Greek first, then have to use Latin. A lot of the elements from the opening "first contact" scenes are derived from Ibn Fadlan's actual account of first encountering the Rus in the 10th century - including the controversial spit bowl washing scene.

One of my favorite scenes in The 13th Warrior is when the Arab ambassadors arrive at the funeral feast and try to communicate with the Northmen in Greek first, then have to use Latin. A lot of the elements from the opening "first contact" scenes are derived from Ibn Fadlan's actual account of first encountering the Rus in the 10th century - including the controversial spit bowl washing scene.

The Last Homely House

26,923 görüntüleme • 4 ay önce

PhD Students – How to identify related papers for your research paper? 1. Go to 2. Paste your DOI or title of your research paper 3. ResearchCollab will display the network of papers 4. Explore the network in following ways → Filter the network of papers based on research field → Filter the network based on thematic clusters → Filter the network based on keywords → and much more As you find related papers, keep saving them. This use case has multiple advantages. ➟ You won’t miss a related paper to cite in your paper ➟ You can identify related papers for your lit review ➟ You can explore new research directions and trends ➟ You can visualize the entire collection of papers As a researcher, work smarter not just harder. For this, try ResearchCollab:

PhD Students – How to identify related papers for your research paper? 1. Go to 2. Paste your DOI or title of your research paper 3. ResearchCollab will display the network of papers 4. Explore the network in following ways → Filter the network of papers based on research field → Filter the network based on thematic clusters → Filter the network based on keywords → and much more As you find related papers, keep saving them. This use case has multiple advantages. ➟ You won’t miss a related paper to cite in your paper ➟ You can identify related papers for your lit review ➟ You can explore new research directions and trends ➟ You can visualize the entire collection of papers As a researcher, work smarter not just harder. For this, try ResearchCollab:

Faheem Ullah

47,305 görüntüleme • 4 ay önce

For robots to be useful, they must be able to interact with a wide variety of environments; and yet, scaling interaction data is difficult, expensive, and time consuming. Instead, much research revolves around sim-to-real manipulation — but mostly this has not been mobile manipulation. Recently, though, this has begun to change. Two recent papers from Tairan He and Haoru Xue show us how to unlock the potential of this technique, building policies which, without any real data at all, can move objects around in the world and open doors in the real world with a humanoid robot. Watch Episode #60 of RoboPapers now to learn more, hosted by Chris Paxton and Jiafei Duan. In this episode, we cover two papers:. First is VIRAL: Visual Sim-to-Real at Scale for Humanoid Loco-Manipulation; and second is DoorMan: Opening the Sim-to-Real Door for Humanoid Pixel-to-Action Policy Transfer.

For robots to be useful, they must be able to interact with a wide variety of environments; and yet, scaling interaction data is difficult, expensive, and time consuming. Instead, much research revolves around sim-to-real manipulation — but mostly this has not been mobile manipulation. Recently, though, this has begun to change. Two recent papers from Tairan He and Haoru Xue show us how to unlock the potential of this technique, building policies which, without any real data at all, can move objects around in the world and open doors in the real world with a humanoid robot. Watch Episode #60 of RoboPapers now to learn more, hosted by Chris Paxton and Jiafei Duan. In this episode, we cover two papers:. First is VIRAL: Visual Sim-to-Real at Scale for Humanoid Loco-Manipulation; and second is DoorMan: Opening the Sim-to-Real Door for Humanoid Pixel-to-Action Policy Transfer.

RoboPapers

30,767 görüntüleme • 9 ay önce

THE WALL BREAKS NEW GROUND! Martin Schindler is through to the knockout stages of the Grand Slam for the first time in his career! The German number one defeats Alexis Toylo 5-2 in their decisive Group C tie! 📺 #GSOD | Group C

THE WALL BREAKS NEW GROUND! Martin Schindler is through to the knockout stages of the Grand Slam for the first time in his career! The German number one defeats Alexis Toylo 5-2 in their decisive Group C tie! 📺 #GSOD | Group C

PDC Darts

37,121 görüntüleme • 11 ay önce

One of the most exciting new AI companies is Sakana AI in our opinion. Their CTO and their Co-founder Llion Jones Llion Jones is now saying it's time to move beyond transformers even though he was one of the 8 original inventors at Google. They are now investigating the next significant step forwards, and have a NeurIPS 2025 spotlight paper called "continuous thought machines" (CTM) which may well be just that (first author Luke Darlow). They just landed their Series B, with the legendary hardmaru as their CEO. They are actively investigating neuroevolution approaches which I strongly believe will be a big part of the future of AI -- and they publish their research in the open. Keep an eye on Sakana! 👌 Interview with Llion and Luke dropping today on MLST.

One of the most exciting new AI companies is Sakana AI in our opinion. Their CTO and their Co-founder Llion Jones Llion Jones is now saying it's time to move beyond transformers even though he was one of the 8 original inventors at Google. They are now investigating the next significant step forwards, and have a NeurIPS 2025 spotlight paper called "continuous thought machines" (CTM) which may well be just that (first author Luke Darlow). They just landed their Series B, with the legendary hardmaru as their CEO. They are actively investigating neuroevolution approaches which I strongly believe will be a big part of the future of AI -- and they publish their research in the open. Keep an eye on Sakana! 👌 Interview with Llion and Luke dropping today on MLST.

Machine Learning Street Talk

187,458 görüntüleme • 11 ay önce

The Talking Equality speaker series is an opportunity to tune in and hear from Thought leaders and Equality champions driving change at Salesforce and beyond. We’re excited to share this experience with you month in and month out. 👋 This Episode — Episode 0, features a behind-the-scenes look at the Business Resource Group program from the perspective of company program managers.

The Talking Equality speaker series is an opportunity to tune in and hear from Thought leaders and Equality champions driving change at Salesforce and beyond. We’re excited to share this experience with you month in and month out. 👋 This Episode — Episode 0, features a behind-the-scenes look at the Business Resource Group program from the perspective of company program managers.

Office of Equality

1,865,981 görüntüleme • 3 yıl önce

It's official — Wispr Flow is now live on the iPhone App Store! We built the first immersive voice keyboard that lets you dictate with incredible accuracy anywhere — 5x faster than typing. Delightful. Effortless. Intelligent. Our mission is to change how people interact with technology and move people from display-first to voice-first interfaces. This product gets us one step closer to empowering people's voices wherever they are.

It's official — Wispr Flow is now live on the iPhone App Store! We built the first immersive voice keyboard that lets you dictate with incredible accuracy anywhere — 5x faster than typing. Delightful. Effortless. Intelligent. Our mission is to change how people interact with technology and move people from display-first to voice-first interfaces. This product gets us one step closer to empowering people's voices wherever they are.

Wispr Flow

130,442 görüntüleme • 1 yıl önce

Happy to share that my PhD paper "Scene2Hap: Generating Scene-Wide Haptics for VR from Scene Context with Multimodal LLMs" has received a Best Paper Award (top < 1%) out of 6,730 submissions at ACM CHI (ACM CHI Conference), the most prestigious conference in the human-computer interaction field🙌 Scene2Hap is an LLM-centered system that automatically designs object-level vibrotactile feedback for entire VR scenes based on objects' semantic attributes (e.g., whether and how the object vibrates) and physical context (e.g., the object's density, spatial relationships). It then renders real-time haptic feedback across the scene, calculating vibration propagation based on LLM-inferred material properties. To the best of our knowledge, this is the first paper to address the problem itself: "designing haptic characteristics of a whole VR scene with one click." Thanks a lot to my co-first-author Easa, Sara Safaee, Paul Strohmeier, and my advisor Jürgen Steimle!

Happy to share that my PhD paper "Scene2Hap: Generating Scene-Wide Haptics for VR from Scene Context with Multimodal LLMs" has received a Best Paper Award (top < 1%) out of 6,730 submissions at ACM CHI (ACM CHI Conference), the most prestigious conference in the human-computer interaction field🙌 Scene2Hap is an LLM-centered system that automatically designs object-level vibrotactile feedback for entire VR scenes based on objects' semantic attributes (e.g., whether and how the object vibrates) and physical context (e.g., the object's density, spatial relationships). It then renders real-time haptic feedback across the scene, calculating vibration propagation based on LLM-inferred material properties. To the best of our knowledge, this is the first paper to address the problem itself: "designing haptic characteristics of a whole VR scene with one click." Thanks a lot to my co-first-author Easa, Sara Safaee, Paul Strohmeier, and my advisor Jürgen Steimle!

Arata Jingu

32,022 görüntüleme • 7 ay önce