Pinned Tweet
VLAs can imitate human demonstrations, but what happens when a task requires a new skill the robot has never seen?
Introducing InSight ๐ค๐ก: self-guided skill acquisition via steerable VLAs! insight-vla.github.io/ ๐งต๐ย [1/N]
Maggie Wang retweeted
Ever wondered what itโs like to learn robot autonomy @Stanford? Now you can.
Today, weโre opening up the robot autonomy teaching materials that my collaborators and I have been developing over the past several yearsโfor anyone, anywhere, completely free.
At the heart of this effort is Principles of Robot Autonomy (PoRA), a textbook inspired by the Stanford robot autonomy course series Iโve been teaching since 2017.
Robot autonomy has evolved from a research discipline into a technology deployed in the real world. Building capable autonomous systems requires more than understanding individual algorithms: it requires understanding how perception, state estimation, planning, control, and learning work together in a complete autonomy stack.
Thatโs what we set out to teach with PoRA: bringing classical robotics and modern AI together in one coherent framework, grounded in both theory and practice.
Coming soon in print from Cambridge University Press, Principles of Robot Autonomy is also freely available online and comes with a comprehensive collection of open educational resources:
๐ Textbook, arxiv.org/abs/2608.03496 โ Foundations of robot autonomy, from classical methods to modern AI-based approaches.
๐ฅ Video lectures, porabook.com/lectures/ Lectures from my Stanford courses on robot autonomy and on optimal and learning-based control.
๐ป Interactive exercises, porabook.com/exercises/ โ Hands-on Python implementations and Jupyter notebooks.
๐ Teaching materials, porabook.com/instructors/ โ Lecture slides and resources for instructors.
๐ Explore all PoRA resources, porabook.com/ โ Including Ask the Book, an AI study assistant integrated throughout the textbook to answer questions, explain derivations, and connect ideas across chapters.
๐ฃ Read the Stanford announcement: news.stanford.edu/stories/20โฆ
Our goal is simple: to make high-quality education in robot autonomy accessible to students, researchers, engineers, educators, and anyone curious about how autonomous robotic systems work.
I believe that sharing knowledge is fundamental to the mission of a university. Autonomous roboticsโand Physical AI more broadlyโare among the defining technological frontiers of our time. We want to help more people understand these technologies, build with them, and advance the field.
A huge thank you to my co-authors, @DanieleGammelli, Joseph Lorenzetti, @katielulula, and @GioeleZardini, and to everyone who helped bring this project to life.
Explore the resources, use them in your own learning or teaching, and please share them with anyone who might find them useful!
#Robotics #PhysicalAI #ArtificialIntelligence #AutonomousSystems #Education
@Stanford @StanfordEng @StanfordAILab @StanfordHAI
Maggie Wang retweeted
๐SteerAct: Stanford ร MIT x Harvard
๐ง 24 test-time rollouts/task. Only steering one feature. ๐Frozen VLA/WAM.
๐คReal-robot mean success: 31.1% โ 52.2% (+21.1 pp; ~68% relative).
๐ฌSee it in action: arxiv.org/pdf/2610.25632
@MIT @medialab @Harvard @Stanford
#Robot #AI
[11/N] Thanks to my incredible collaborators: Lars Osterberg, @stephentian_, Ola Shorinwa, @jiajunwu_cs, and @MacSchwager for making this work possible! ๐๐
1
1
413
[12/N] Check out the links below! ๐
๐ Project page: insight-vla.github.io
๐ Paper: arxiv.org/abs/2606.24884
๐ฉ๐ปโ๐ป Code: github.com/insight-vla/insigโฆ
2
390
๐ค Excited to announce our Workshop on Generative Digital Twins for Real2Sim & Sim2Real Transfer in Robotics at ICRA 2026 in Vienna! ๐ฆ๐น
How can generative digital twins help bridge the persistent gaps between reality โ simulation (Real2Sim) and simulation โ reality (Sim2Real) in robotics?
๐๏ธ Workshop: June 1, 2026
๐ Submission deadline: March 20, 2026
๐ awesomedigitaltwin.github.ioโฆ
๐ก Speakers: @IngmarPosner (Oxford), @oier_mees (Microsoft), @yukez (NVIDIA / UT Austin), @jiajunwu_cs (Stanford), @ManolisSavva (SFU), @HengshuangZhao (HKU), @bgxc (Lightwheel)
๐ Survey: arxiv.org/abs/2504.13159
๐ธ๏ธ SOTA mind map: miro.com/app/board/uXjVI6nHEโฆ
๐ GitHub: github.com/ndrwmlnk/awesome-โฆ
Org team: Andrew Melnik, @longhini_a, Jonathan Francis, @UksangYoo, @liuzhixuan0704, @maggi3wang, Mehreen Naeem
#ICRA2026 #DigitalTwins #Sim2Real #Real2Sim #GenerativeAI #Robotics #3DReconstruction
Humans estimate object physical properties through vision and refine them through touch. ๐๐ข๐ฏ ๐ณ๐ฐ๐ฃ๐ฐ๐ต๐ด ๐ฅ๐ฐ ๐ต๐ฉ๐ฆ ๐ด๐ข๐ฎ๐ฆ?
Excited to share ๐ฃ๐ต๐๐2๐ฅ๐ฒ๐ฎ๐น ๐ค๐๏ธโ: VLM priors + interactive adaptation for uncertainty-aware sim-to-real! phys2real.github.io/ ๐งต (1/N)
(9/N) This was a fun collaboration with @stephentian_, @SwannAiden, Ola Shorinwa, and my amazing advisors @MacSchwager and @jiajunwu_cs! ๐
1
2
685
Check out the project website ๐: phys2real.github.io/
Paper ๐: arxiv.org/abs/2510.11689
3
619




