How can we reconstruct 3D objects under *extreme lighting... variations*? 🌤️🌥️🌆🌃🌉 How about ... 🤔 appearance embedding? BUT it cannot capture view-dependent appearances. 🤔 inverse rendering? BUT it suffers from ambiguities. 💡 Our idea: Relighting comes to the rescue!show more

But what about hair? Dynamic 3D Gaussian Splatting can... impressively reconstruct it. I find it fascinating how accurately a bunch of ellipsoids can capture the complexity of hair. Amazing!show more

Omg 🙈 The camera man’s worried lol someone told... him about the post and he’s asked to come round and see me for a chat” But I think I can handle it 🤔I’ll let you know how it goes 🫣show more

COVER 6 is trending right now, but how do... you play it? Everyone knows what it looks like on paper, but how does it play out on the field? This is a great play to study. How are you coaching it? Who gets a plus and who gets a minus? I've sent this to 3 different NFL DCs—they all gave different answers!! Curious to see how others view it. I'll post my breakdown later. Coaching is about details, but which details? Remember, We cannot give players complex algorithms to solve - rule and keys need to be simple and repeatable to hold up under pressure and unscripted looks. Assign 3 players most responsible for this play 1,2,3. This is a fun 1!!show more

“Black Walker” dogman in motion...🧐🤔👹👹 Apparently this video was... taken in Smuggler’s Notch, VT, back in 2009. I’ve seen it claimed as coming from a few different sources, but as you can see from the video, this is NOT just “look at that obvious AI slop!!” but genuine wildlife footage worthy of study. What do you make of its gait, appearance, and how it interacts within the environment?show more

Introducing Kaleido💮 from AI at Meta — a universal... generative neural rendering engine for photorealistic, unified object and scene view synthesis. Kaleido is built on a simple but powerful design philosophy: 3D perception is a form of visual common sense. Following this idea, we formulate rendering purely as a sequence-to-sequence generation problem, successfully unifying neural rendering with the architecture principles behind modern language and video models. Unlike traditional neural rendering methods, Kaleido learns 3D purely in a data-driven way, without explicit 3D representations or structures. It acquires spatial understanding directly through large-scale video pretraining, then multi-view 3D data finetuning, inspired by how LLMs acquire textual common sense from large corpora before specialising in domains like coding. Through extensive ablations, we progressively modernised the architecture design and training strategies and tackled key scaling challenges in sequence-to-sequence generative rendering, arriving at a design that’s simple, versatile, and scalable. Kaleido significantly outperforms prior generative models in few-view settings, and remarkably is the first zero-shot generative method matches InstantNGP-level rendering quality in multi-view settings. We view Kaleido also as an alternative step towards world modeling that flexibly spans a spectrum of “realities": with many views, it faithfully reconstructs grounded reality; with fewer views, it imagines plausible unseen details. 🔗 Explore more results and paper:show more

At the ruins of the ancient Quarries of Cusa... also known as Cave di Cusa, in Italy, we see limestone cylindrical columns ranging from over 6 feet in diameter to 12 feet! Modern eyes would instantly relate how they were cut to heavy machinery. But they obviously didn’t have such equipment, so how did they do it!? 🤔show more

Two months ago, at an RSS workshop panel, I... said that general real2sim reconstruction is still far from solved, especially for articulated objects, which will be a major bottleneck for robotics. But now that has totally changed with Astra; using several photos as input, we can easily reconstruct vivid articulated scenes. The result is so impressive! Excited to see how it will shift the paradigm of real2sim2real and agentic learning for robotics and humanoids.show more

I can teach people how to pronounce “Chagee” and... “Mixue” sebab cina ma, but when it comes to CHA-TRA-MUE the Thai tea brand, I need help from the locals hahahahahaha so now we know, “MUE” is not “mwe” or “miu”, it’s like ME in MEnang!show more

Most chiral molecules arise from carbons being bonded to... 4 different atoms, which are called sterocenters. The makes the molecule have a different mirror image that cannot arise from simple rotation. But, you can have chiral molecules not from stereocenters. You can have chirality that doesn't come from a single point in the molecule. It comes from some global property. The classic example is helicene, which doesn't have any stereocenters, but has chirality because of hits helical structure. This means you cannot capture this molecule with a graph, and thus SMILES or a string representation cannot capture this. Of course natural language comes to the rescue (just say in words if it's left-handed or right-handed helix), but it's an interesting failure mode for viewing molecules as just a graph. Another example of a molecule with helical chirality is DNA. DNA is actually chiral in two ways, which is kind of confusing. It has both helical structure and stereocenters. You won't find the stereocenters ever flipped, but left-handed helical DNA can exist (called Z-DNA). Interestingly, making the flipped stereocenter of DNA could be part of an entire mirror organism (mirror RNA, DNA, AAs, sugars) that would then be potentially invisible to our immune systems. This has been recently proposed as a "mechanism" for how a runaway AI system could cause harm to Earth. I find it to be a pretty tedious and difficult way to cause harm, but it is intellectually cool. Anyway - this came up in a PhD defense and I have a lot of arcane knowledge about this I wanted to dump.show more

As a mature adult, we should make a decision... when we face two or more options. Before we make decision, we can spend some time to question it and do a lot of research. But after we make decision, we must try our best to work on our goal and contribute to it. If we feel it is not going as we expected or we don't want to study how to improve it. We can make a decision to stop it instead of complaining. Because complaining cannot resolve anything. It will bring a lot of negative energy in our body. If we spread our complaint to the community or to put family or our colleagues , we can only let them feel we are very weak who always need attention and we are not a good person to cooperate or get along. When we in Pi Network, complaint cannot help but probably bring bad luck. If we don't like it, we have an option to quit it. But if we don't want to quit we should study and catch up! Community don't need a negative voice. Community need pioneers who knows what is appreciation and actively participate to empower Pi Network! Look at the following four videos. These pioneers are good examples! Pi Network #WhatIDoForPi #PiGCV #Pioneershow more

truth will come out no matter how much you... try to hide it. and the good thing about karma, is, that it comes back— good & bad 🙇‍♀️ them who exude nothing but authenticity & kindness. you can never put these pure souls down. we will always stand with, by, and for donny & belle. 🫂show more

honestly, i'm uncomfortable with how every bts content gets... h3rassed like this by kpop fans, especially on tiktok. what can we do about tiktok? and what's even scarier is that the majority participating in this bts hate train are under 18, but people don't care about it. most of you might not realize, but bts has antis just as dedicated as their loyal fans. this isn't about fanwars, we need to unite to stop the spread of misinformation…show more

Jay Johnson was great to carve out some time... in his day to chat with us, but we all knew that when the phone rang my guy was bouncing. As fate would have it, just about four minutes into our convo, he got a phone call from a recruit. Not sure who it was or how it went, but Jay was pretty pleased when he returned. #LSUshow more

Gaussian Head Avatar: Ultra High-fidelity Head Avatar via Dynamic... Gaussians paper page: Creating high-fidelity 3D head avatars has always been a research hotspot, but there remains a great challenge under lightweight sparse view setups. In this paper, we propose Gaussian Head Avatar represented by controllable 3D Gaussians for high-fidelity head avatar modeling. We optimize the neutral 3D Gaussians and a fully learned MLP-based deformation field to capture complex expressions. The two parts benefit each other, thereby our method can model fine-grained dynamic details while ensuring expression accuracy. Furthermore, we devise a well-designed geometry-guided initialization strategy based on implicit SDF and Deep Marching Tetrahedra for the stability and convergence of the training procedure. Experiments show our approach outperforms other state-of-the-art sparse-view methods, achieving ultra high-fidelity rendering quality at 2K resolution even under exaggerated expressions.show more