Xiaolong Wang
San Diego, California, United States
8K followers
500+ connections
About
I am an assistant professor at UC San Diego in the ECE department. I am affiliated with…
Activity
-
Today was my last day at Wayve. Thank you Alex Kendall and Jamie Shotton for the opportunity to work at the very cutting edge of embodied AI…
Today was my last day at Wayve. Thank you Alex Kendall and Jamie Shotton for the opportunity to work at the very cutting edge of embodied AI…
Liked by Xiaolong Wang
-
It's official. I was awarded the IEEE RAS George Saridis Leadership Award in Robotics and Automation. To be presented at the IEEE RAS ICRA conference…
It's official. I was awarded the IEEE RAS George Saridis Leadership Award in Robotics and Automation. To be presented at the IEEE RAS ICRA conference…
Liked by Xiaolong Wang
-
Congratulations to TILOS Robotics team co-lead Henrik Christensen and team member Xiaolong Wang! Dr. Christensen has been recognized with the 2026…
Congratulations to TILOS Robotics team co-lead Henrik Christensen and team member Xiaolong Wang! Dr. Christensen has been recognized with the 2026…
Liked by Xiaolong Wang
Experience
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
-
Education
-
-
-
-
-
-
-
Honors & Awards
-
Baidu Scholarship
Baidu
http://scholarship.baidu.com/
-
Facebook Fellowship
Facebook
https://research.fb.com/fellows/wang-xiaolong/
-
NVIDIA Fellowship
NVIDIA
https://research.nvidia.com/grad_fellows/2017/XiaolongWang
More activity by Xiaolong
-
Excited to share a couple of milestones from today. I successfully passed my PhD thesis proposal, “Autonomy for Heterogeneous UGV-UAV Fleets,”…
Excited to share a couple of milestones from today. I successfully passed my PhD thesis proposal, “Autonomy for Heterogeneous UGV-UAV Fleets,”…
Liked by Xiaolong Wang
-
Grateful to share that I’ve been promoted to Senior Director, Robotics Software Engineering at NVIDIA. Last week also marked my 5th year here -…
Grateful to share that I’ve been promoted to Senior Director, Robotics Software Engineering at NVIDIA. Last week also marked my 5th year here -…
Liked by Xiaolong Wang
-
New TRIP-Bag is here! TRIP-Bag (Teleoperation, Recording, Intelligence in a Portable Bag) is KIMLAB (Kinetic Intelligent Machine LAB)’s new robot…
New TRIP-Bag is here! TRIP-Bag (Teleoperation, Recording, Intelligence in a Portable Bag) is KIMLAB (Kinetic Intelligent Machine LAB)’s new robot…
Liked by Xiaolong Wang
-
Zipping a bag was the first two-hand coordination task we tried 2 yrs ago. It didn’t work back then, and we were bummed. But excited now that it…
Zipping a bag was the first two-hand coordination task we tried 2 yrs ago. It didn’t work back then, and we were bummed. But excited now that it…
Liked by Xiaolong Wang
-
Thank you, Singapore! Wrapping up a part-time position in Singapore this weekend. I have been 50% with Temasek Holdings since April 1, 2025. My…
Thank you, Singapore! Wrapping up a part-time position in Singapore this weekend. I have been 50% with Temasek Holdings since April 1, 2025. My…
Liked by Xiaolong Wang
-
I’m excited to share that I’ve decided to join the Microsoft SuperIntelligence team. I believe this is an opportunity to work on something that goes…
I’m excited to share that I’ve decided to join the Microsoft SuperIntelligence team. I believe this is an opportunity to work on something that goes…
Liked by Xiaolong Wang
-
Introducing EgoVerse: an ecosystem for robot learning from egocentric human data. Built across 4 academic labs and 3 industry partners, EgoVerse is…
Introducing EgoVerse: an ecosystem for robot learning from egocentric human data. Built across 4 academic labs and 3 industry partners, EgoVerse is…
Liked by Xiaolong Wang
-
Congratulations to Nikhil Shinde! When we first started working together we spent many hours over zoom during COVID trying to figure out good…
Congratulations to Nikhil Shinde! When we first started working together we spent many hours over zoom during COVID trying to figure out good…
Liked by Xiaolong Wang
-
Can a single reward model work zero-shot across robots, tasks, and environments? We’re excited to introduce Robometer, a general-purpose robotic…
Can a single reward model work zero-shot across robots, tasks, and environments? We’re excited to introduce Robometer, a general-purpose robotic…
Liked by Xiaolong Wang
-
Here's the official flyer for our IEEE RoboSoft 2026 workshop, taking place on April 7th, 2026, in Kanazawa, Japan! Please join us and consider…
Here's the official flyer for our IEEE RoboSoft 2026 workshop, taking place on April 7th, 2026, in Kanazawa, Japan! Please join us and consider…
Liked by Xiaolong Wang
-
Reve's new text-to-image model is here. Really proud of the team to rank at #3 with the big labs. To my knowledge, we are the first lab to use native…
Reve's new text-to-image model is here. Really proud of the team to rank at #3 with the big labs. To my knowledge, we are the first lab to use native…
Liked by Xiaolong Wang
-
We release Cosmos Policy: a state-of-the-art robot policy built on a video diffusion model backbone. - policy + world model + value function — in 1…
We release Cosmos Policy: a state-of-the-art robot policy built on a video diffusion model backbone. - policy + world model + value function — in 1…
Liked by Xiaolong Wang
-
SR-3D was accepted to #ICLR2026, and the code has now been released! We also release SR-3D-Bench, a region-level spatial QA benchmark on egocentric…
SR-3D was accepted to #ICLR2026, and the code has now been released! We also release SR-3D-Bench, a region-level spatial QA benchmark on egocentric…
Liked by Xiaolong Wang
-
I have 150+ apps on my phone because I'm always trying things as an investor, and Alta has remained one of my daily favorites. Congrats to Jenny…
I have 150+ apps on my phone because I'm always trying things as an investor, and Alta has remained one of my daily favorites. Congrats to Jenny…
Liked by Xiaolong Wang
Other similar profiles
Explore more posts
-
S A M Matiur Rahman
Google’s groundbreaking Gemini Robotics 1.5 propels Artificial Intelligence (AI) from code to concrete, empowering robots to reason, plan, and seamlessly collaborate with humans in the physical world. By integrating advanced digital tools and cross-embodiment learning, these AI agents translate complex instructions into fluid, purposeful movements with unprecedented precision. The dual-model architecture—Gemini Robotics-ER 1.5 as the strategic “brain” and Gemini Robotics 1.5 as the agile “body”—unlocks a new era of adaptive, intelligent automation. This milestone not only redefines robotics but also charts a bold course toward practical AGI that transforms how we live, work, and innovate. 💚 #GeminiRobotics #AIInnovation #RoboticsRevolution #AGI
-
Louis Nguyen
Koopman Operators in Robotics just received a major 2026 update. A new IEEE Transactions on Robotics review (published a few days ago) summarizes where Koopman-based robot learning stands today: global linearization, runtime learning, and high-dimensional lifting for complex robots. I read the full 20-page review, and one thing became clear: Koopman (2026) is powerful — but it stops at fixed admissibility. Once the environment changes, constraints shift, contact becomes discontinuous, uncertainty becomes multi-modal, or the regime changes… the operator structure breaks. This is exactly what the authors list in Section VI: - Constraint lifting remains unsolved - Stochasticity breaks the lifted space - Sampling rate changes the operator - Contact dynamics create discontinuities - Soft robots become intractable - Hybrid systems break invariance - Uncertainty is not preserved under lifting - Optimization feasibility is not guaranteed These are not weaknesses — they are structural limits of any framework that assumes a fixed state space and external clock time. And this is where my work over the past year naturally diverged, built from the opposite direction: Start with admissibility → define geometry → attach time → then reconstruct valid operators without lifting. When admissibility evolves, the state space evolves. When the geometry shifts, time itself shifts with it. Regime transitions, hazard-driven collapse, and recovery are first-class geometric events — not errors, not noise, not discontinuities. To visualize this, I created a small PNG comparing “Where Koopman (2026) stops” vs “Where Hazard Manifold + VORT begin.” It’s not competition — it’s simply two paradigms built on different first principles. For me, this 2026 review is a milestone: the robotics community has now mapped the boundary of the Koopman paradigm, just as admissibility-first geometry is beginning to take shape. It’s exciting to see both worlds advancing. One stabilizing a powerful operator-based approach. The other rebuilding dynamics from existence and admissibility upward. Onward. Louis Nguyen 2026
-
Robotics MDPI
🔓 New paper online "Optimizing Coalition Formation Strategies for Scalable #MultiRobot #TaskAllocation: A Comprehensive Survey of Methods and Mechanisms" in Robotics MDPI, by Krishna Arjun, David Parlevliet, Hai Wang and Amirmehdi Yazdani from Murdoch University. 👉 Welcome to read: https://lnkd.in/gvii_-8w MDPI
-
SAI QIAN ZHANG
✨ Excited to share VisArch 2026, the ASPLOS 2026 Workshop on Systems and Architectures for Neural Rendering, AR/VR, and Visual Computing, taking place on March 22, 2026 in Pittsburgh, PA. VisArch brings together researchers across computer architecture, AI, XR, neural rendering, and visual computing to discuss efficient, low-latency, and scalable platforms for next-generation visual workloads. 🎤 The program will feature keynote talks from Sarita Adve (UIUC), David Lindlbauer (CMU), and Vincent T. Lee (Meta), along with poster presentations on topics such as 3D Gaussian Splatting, SLAM, efficient XR systems, and contextual visual intelligence. Looking forward to a day of engaging discussions with the community at the intersection of systems, architecture, and visual computing. 🔗 Learn more here: https://lnkd.in/eF2d6hbi #ASPLOS2026 #VisArch #ARVR #XR #NeuralRendering #VisualComputing #ComputerArchitecture
-
Junwei Liang
Our CoRL 2025 paper, GLOVER++, tackles a key question: How can we leverage human videos to boost robot manipulation performance and generalization? 🤖 We're excited to release HOVA-500K, a dataset of 500K annotated human object interaction images! It's perfect for training affordance reasoning models (e.g., "pour me water": given an image, the model outputs the mug handle's location). This enables zero-shot robot manipulation and enhances imitation learning precision. 🙌 Code & dataset are open-source! Check out GLOVER++! 👇 https://lnkd.in/gdUi7fQY #CoRL2025 #Robotics #AI
-
Pakpong Chirarattananon
Happy to see our Perspective now out in npj Robotics on a performance-driven approach for soft robotics. https://lnkd.in/e2466K2Q From our side, some of this thinking has been shaped by work on the Hopcopter—where compliance isn’t just for safety, but a way to store and redirect energy for impulsive motion. That experience really convinced me that soft elements can increase performance when thoughtfully integrated. Of course, this only works in synergy: compliant mechanisms paired with rigid structures (beautifully explored in Kaushik Jayaram’s work) and emerging high-performance soft actuators (Kevin Chen and others) are what make these hybrid systems compelling. If this resonates, we’re organizing a RoboSoft 2026 workshop on Fast and Impulsive Robots Enabled by Soft Actuators and Mechanisms (April 7, Kanazawa). Would love to continue the conversation there. https://lnkd.in/eD2QWJuE
-
Simon Daubner
"evoxels: Differentiable Physics for Voxel-Based Microstructure Simulation" is now published in the Journal of Open Source Software🥳🥳 After a couple months of waiting and a genuinely thoughtful, constructive review process on GitHub, it’s wonderful to see it out in the world. Huge thanks to David Zwicker, Nils Meyer and Philip Cardiff. evoxels is a differentiable physics framework for voxel-based microstructure simulations, designed to connect imaging, simulation, inverse modeling, and machine learning in a unified open-source workflow. Big shout out to my amazing collaborators Alexander Cohen, Benjamin Dörich and Sam Cooper! I’m also really excited that there’s more to come soon, with new timestepping methods, multiphase-field simulations, and other improvements already in the pipeline. Updates soon 👀 And for all nerds out there who love language beyond coding, here's a little treat inspired by the Jungle Book by Rudyard Kipling: Law of the Micros Now this is the Law of the Micros, determined by fine structure scale; Those mastering it shall prosper, solvers that don’t become stale. Ions move between electrodes, differentiable physics go forward and back For the strength of the mesh is the voxel, and the strength of the voxel its pack. Model just as complex as it must be; use data, but only what's FAIR; For powerful compute is frugal - be mindful, resourceful, aware. Brute-force may chase after fortune, but friend, as your insight has grown, Remember: the solver's your hunter. Extract hidden truths of your own. Keep pace with the law of the physics - the transport, reactions, and flow. And honour the scale of the micro, for structure decides how things go. #MaterialsScience #Simulation #MachineLearning #OpenSource #Python
5 Comments -
Khaza Anuarul Hoque
Our Lab has made significant strides in the field of safe, secure, and energy-efficient AI hardware this year. Below are key developments in AI/GenAI hardware research: 1. **XAI-Gen Methodology**: We introduced a novel approach leveraging emerging hardware accelerator models (e.g., Google TPU v4) and explainable AI (XAI) to identify non-critical layers for approximation in approximate deep neural networks (AxDNNs). Our findings indicate that XAI-Gen can reduce energy consumption by up to 7× with only a 1-2% accuracy loss. The publication is available here: https://lnkd.in/gy9RRuvt. 2. **Adversarial Robustness and Hardware Efficiency**: Building on our XAI-Gen framework, we co-optimized adversarial robustness and hardware efficiency for AxDNNs on systolic-array hardware accelerators. The results were published in the IEEE Transactions on Sustainable Computing (2025): https://lnkd.in/ggrGBp7w. 3. **Accelerating XAI Algorithms**: We explored methods to enhance the efficiency of XAI algorithms on AI hardware through our XAIedge framework. By integrating approximate computing techniques into XAI algorithms—such as integrated gradients and Shapley analysis—XAIedge translates these algorithms into approximate matrix computations. It achieves up to 2× improvement in energy efficiency compared to existing methods while maintaining comparable accuracy. The research is published in IJCNN 2025, with the preprint available here: https://lnkd.in/gNYDEMD2. 4. **Fault Detection in AxDNNs**: Recognizing the difficulty of detecting faults in AxDNN hardware, we introduced a lightweight framework named EPSILON. By utilizing pre-computed statistical signatures and layer-wise importance metrics, EPSILON efficiently identifies and mitigates faults, outperforming existing detection methods. This work is also published in IJCNN 2025: https://lnkd.in/gtRkPX5Z. 5. **Hardware Trojan Detection**: Our recent paper investigates whether large language models (LLMs) can assist in hardware Trojan detection. We propose a detection and explanation method using pre-trained LLMs, including enterprise API-enabled models like GPT-3.5 Turbo and Google Gemini 1.5 Pro, as well as open-source models such as Meta AI Llama-3.1. Our approach emphasizes transparency and trust in the IC supply chain and is published in IEEE Access (2025): https://lnkd.in/gMEyxap8 6. **LLM Compression with TOGGLE**: We addressed the feasibility of preserving LLM behavior post-compression, introducing TOGGLE—an innovative framework that incorporates formal methods into LLM compression. Evaluating TOGGLE on four architectures (GPT-2, DeepSeek-V2 7B, LLaMA 3 8B, and Mistral 7B), we achieved up to 3.3× reduction in computational costs and a 68.8% decrease in model size while maintaining linguistic integrity. This work has been recently accepted for ICCAD 2025. #GenAI #AI #Hardware #EdgeAI #EDA
13 Comments
Explore top content on LinkedIn
Find curated posts and insights for relevant topics all in one place.
View top content