We're hiring in robot learning at @theworldlabs! Join me, @drfeifei, and the team to define and scale the next generation of world models for robot learning! Atlas for Robotics: worldlabs.ai/blog/atlas#robo… Real-to-Sim-to-Real: worldlabs.ai/blog/real-to-si… Apply: jobs.ashbyhq.com/worldlabs/8…

Naively finetuning a generalist policy on a new task often just turns it into a specialist: it forgets all the tasks it was trained on before. 🥲 This happens regardless of whether the base policy is large or small (e.g., VLA or diffusion policy). So, how to actually get the continual learning without forgetting? You need to find the "memory anchors" and train with them! Check out @du_maximilian's 🧵👇

Few-shot, in-context learning for robotics is here. Super exciting! Unlike GPT-3, whose pretraining data largely came from the Internet, GEN-1.5 was trained on large-scale, deliberately collected physical interaction data. A strong signal that scaling real-world interaction data can unlock new capabilities in robotics.

Agentic Real2Sim is an interesting take on world modeling. Instead of learning physics directly from videos, it lets a Vision-Language Agent reconstruct a physics simulator from a single real-world video. The result is an editable, executable world model that can be queried, simulated, and improved.

Congratulations to UBC Computer Science Assistant Professor Peter Yichen Chen and CS Associate Member Chao Liu for their Outstanding Systems Paper Award at the Robotics: Science and Systems (RSS) conference!

🚀 Excited to kick off the Graphics × Science Workshop at #SIGGRAPH2026! Computer graphics is becoming a foundational tool for scientific discovery. From computational imaging and molecular modeling to physical simulation, robotics, manufacturing, and AI, graphics is helping us model, understand, and design the physical world. Looking forward to an exciting program featuring 3 keynote speakers and 58 highlighted papers. 🔗 graphics4science.github.io/2… #AI4Science #ComputerGraphics #ScientificComputing #Simulation #NVIDIA #NVIDIAOmniverse #NVIDIAAI

Grateful to all my mentors, mentees, and collaborators for being part of this journey! Excited to continue pushing the boundaries of simulation for Physical AI.

Excited to share our #SIGGRAPH 2026 Technical Workshop: Differentiable Physics for Graphics and AI Save the date for Monday, July 20: an afternoon of keynotes, lightning talks, and discussion! dpgai.github.io/SIGGRAPH2026… #DifferentiablePhysics #ComputerGraphics #PhysicalAI

Villeroché et al., "Zero-shot generalization of transformer neural operators to larger domains" Rotary Positional Encoding (RoPE) works well, but is non-local, which can make extrapolation hard. You can add locality via a position bias.

Introducing Universal Manipulation Exoskeleton (UME) A low-cost exoskeleton with real-time haptic torque feedback for learning autonomous policies that perform highly force-mediated, tightly space-constrained, visually occluded, whole-body, and long-horizon mobile manipulation tasks. Using UME, the teleoperator can unsheathe a heavy metal sword completely blindfolded. ume-exo.github.io/ 🧵1/N

#CVPR2026 Zero simulation data. Zero annotations. One physics foundation model that generalizes across arbitrary geometries. Introducing PhysSkin: a transformer-based, physics-informed self-supervised model that learns real-time simulation straight from static 3D shapes. The representation is discretization-agnostic — a single model generalizes across object categories, topologies, and resolutions, and works directly on 3D Gaussians. Under the hood: a transformer point-cloud encoder extracts latent shape features, while a cross-attention decoder aggregates both surface and volumetric information. Sun, Jun 7, 2026 · 10:45 AM–12:45 PM PDT · ExHall F 348 zju3dv.github.io/PhysSkin/

Want to know more about our new elastodynamic contact simulator inside Genesis World? See our #SIGGRAPH 2026 paper at simulation-intelligence.gith…! In 2020, IPC enabled penetration-free simulation using barrier functions, but it limited the efficiency. Today, let's remove the barrier!

Introducing ✨RigidFormer: Learning Rigid Dynamics with Transformers - our attempt to scale learning-based physical dynamics with Transformers. RigidFormer learns rigid dynamics with Transformers. It is a mesh-free, object-centric Transformer for multi-object rigid-body contact dynamics from point clouds. Learning physics with purely neural simulators, without relying on traditional physics engines, is an important and widely studied problem. Prior SOTA methods often use graph neural networks for accuracy and generalization, but still struggle with efficient, high-fidelity simulation at scale. RigidFormer uses only point inputs, matches or outperforms mesh-based baselines on standard benchmarks, runs much faster, generalizes across point resolutions and datasets, and scales to 200+ objects. We also show a preliminary extension to command-conditioned articulated bodies by treating body parts as interacting object-level components. RigidFormer is mesh-free: it does not require mesh connectivity, SDFs, or vertex-level message passing, making it well-suited for point-cloud observations and scalable simulation. This architecture can also be adapted to learn soft-body dynamics by replacing the rigid-body module (differentiable Kabsch alignment). 🎬See our video for more details. Many thanks to my amazing collaborators: Minghao Guo @GuoMh14, Haixu Wu @Haixu_Wu_1998, Doug Roble, Tuur Stuyck @TuurStuyck, and Wojciech Matusik @wojmatusik. Project page: people.csail.mit.edu/frankzy… Paper: people.csail.mit.edu/frankzy…

Excited to share that our work NeuralActuator: Neural Actuation Modeling for Robot Dynamics and External Force Perception has been accepted to #RSS2026! Your robot — even a low-cost one — can feel external forces without torque or tactile sensors. TL;DR: NeuralActuator is a neural actuator model that jointly predicts 1️⃣torque to capture the nonlinear and time-varying current–to–torque relationship of low-cost servos, 2️⃣external contact forces (and force detection gates) for sensorless force perception, 3️⃣and motor conditions that indicate each motor’s operating regime. Here is a fast-forward video clip ⬇️ We are also covering more robots like LeRobot-S101 and Franka Panda. More details coming soon.

SAD: Soft Anisotropic Diagrams for Differentiable Image Representation has been accepted by #SIGGRAPH2026 Check it out, and huge congrats to Lucky! @Luckyballa #SAD represents an image as a soft, anisotropic, differentiable diagram over learnable sites. Each pixel is modeled as a softmax blend over its top-K nearby sites under a site-dependent distance, yielding a differentiable partition of unity with explicit ownership and content-aligned boundaries. A GPU-friendly top-K propagation scheme keeps the cost constant per pixel, enabling fast fitting at matched or better quality. Classical geometric structures can still inspire fresh perspectives in modern visual computing. Voronoi and Power diagrams have long been elegant tools for 3D shape analysis, reconstruction, and geometric reasoning; here, related diagram ideas, with connections to Apollonius-style diagrams, are explored for image representations. Homepage: luckyiyi.github.io/SAD/ arXiv: arxiv.org/pdf/2604.21984 #SIGGRAPH2026 #SIGGRAPH #CV #Vision #Graphics #CG

Excited to share GeoPT: Scaling Physics Simulation via Lifted Geometric Pre-Training, which received the Best Paper Award at the ICLR 2026 Workshop on Foundation Models for Science. Can we scale neural physics simulation without scaling expensive solver-generated labels? (1/6) (The below results are all predicted by GeoPT.)

What happens when differentiable simulation meets diffusion models? You get a foundation model that is both 𝗲𝘅𝗽𝗿𝗲𝘀𝘀𝗶𝘃𝗲 AND 𝗴𝘂𝗮𝗿𝗮𝗻𝘁𝗲𝗲𝗱 𝘁𝗼 𝗼𝗯𝗲𝘆 𝗽𝗵𝘆𝘀𝗶𝗰𝗮𝗹 𝗰𝗼𝗻𝘀𝘁𝗿𝗮𝗶𝗻𝘁𝘀. 📢 Excited to share our latest work accepted to #ICLR2026: "Physically Valid Biomolecular Interaction Modeling with Gauss-Seidel Projection." 🧬 Foundation models like AlphaFold3 and Boltz have transformed biomolecular structure prediction — yet they still hallucinate physically invalid structures: steric clashes, distorted covalent geometry, broken stereochemistry. The root cause? Current predictors are trained to match empirical distributions — they never enforce physical validity as a 𝗵𝗮𝗿𝗱 𝗰𝗼𝗻𝘀𝘁𝗿𝗮𝗶𝗻𝘁. The secret sauce? A 𝗱𝗶𝗳𝗳𝗲𝗿𝗲𝗻𝘁𝗶𝗮𝗯𝗹𝗲 𝗚𝗮𝘂𝘀𝘀-𝗦𝗲𝗶𝗱𝗲𝗹 𝗽𝗿𝗼𝗷𝗲𝗰𝘁𝗶𝗼𝗻 𝗺𝗼𝗱𝘂𝗹𝗲 that takes provisional atom coordinates from the diffusion model and projects them onto the nearest physically valid configuration. By exploiting the locality and sparsity of atomic constraints, it converges stably and fast at scale. The module plugs into existing frameworks end-to-end via implicit differentiation — treating physical validity as a first-class citizen during 𝗯𝗼𝘁𝗵 𝘁𝗿𝗮𝗶𝗻𝗶𝗻𝗴 𝗮𝗻𝗱 𝗶𝗻𝗳𝗲𝗿𝗲𝗻𝗰𝗲. With our projection module in place, just 𝟮 𝗱𝗲𝗻𝗼𝗶𝘀𝗶𝗻𝗴 𝘀𝘁𝗲𝗽𝘀 suffice to match the structural accuracy of 200-step diffusion baselines — delivering a ~𝟭𝟬× 𝘄𝗮𝗹𝗹-𝗰𝗹𝗼𝗰𝗸 𝘀𝗽𝗲𝗲𝗱𝘂𝗽 while guaranteeing physical validity. Across six benchmarks (CASP15, PoseBusters, AF3-AB, dsDNA, RNA-Protein, and more), our model closes the gap between guaranteed physical validity and state-of-the-art structural accuracy. Joint work across UBC, MIT, NVIDIA, PKU, U of Utah, and Foundry Biosciences. 📄 Project: chensiyuan030105.github.io/P… 💻 Code: github.com/chensiyuan030105/…

Modern all-atom biomolecular foundation models can be impressively accurate, but they still often generate steric clashes and other physically invalid structures. In our ICLR 2026 paper, we ask: can physical validity be enforced as a hard constraint, instead of a soft preference? Many existing approaches use physics mainly as inference-time guidance. That can reduce violations, but with finite guidance strength and finite denoising steps, invalid structures can still slip through. We wanted guarantees. (1/6)

Last year, Graphics4Science started at SIGGRAPH as a course. This year, it returns as a workshop, with a stronger cross-disciplinary focus: Graphics × Science. We are calling for papers to highlight resent research in this direction. Details in below.

UBC Computer Scientists and collaborators show that an AI system can conduct independent research, potentially accelerating research discoveries. Read more in our new article: cs.ubc.ca/news/2026/03/ai-sc…