Ruwan Wickramarachchi (@ruwantw) header
Ruwan Wickramarachchi (@ruwantw) profile picture

Research Scientist @BoschUSA | Ph.D. #AIISC @UofSC | Neurosymbolic AI | Hybrid Modeling | Robotics | Autonomous Systems

Pittsburgh, PA
Joined June 2010
Photo by Ruwan Wickramarachchi (@ruwantw)Photo by Ruwan Wickramarachchi (@ruwantw)Photo by Ruwan Wickramarachchi (@ruwantw)Photo by Ruwan Wickramarachchi (@ruwantw)Photo by Ruwan Wickramarachchi (@ruwantw)Photo by Ruwan Wickramarachchi (@ruwantw)Photo by Ruwan Wickramarachchi (@ruwantw)Photo by Ruwan Wickramarachchi (@ruwantw)Photo by Ruwan Wickramarachchi (@ruwantw)Photo by Ruwan Wickramarachchi (@ruwantw)
Attending #CoRL26 in Austin, TX. We are presenting 3 papers. Main: @Bw_Li1024 is presenting our work on ReSYNC; @Shuangyu0_0Xie and @kaiyuan_c are presenting our work on Graph-as-Policy (GaP) Learn@Deploy workshop: @revvozz is presenting our work on REMM tomorrow at 10:30am.
1
2
10
637
@revvozz is presenting REMM now in the poster session at JW Marriot 2nd floor.
1
1
1
81
Ruwan Wickramarachchi retweeted
Introducing GR00T 2: our most advanced, open System One model for robotics. It runs entirely on your robotโ€™s onboard compute, fully untethered, and you own every weight of it. GR00T 2 features: 1. Long-horizon tasks: dexterous manipulation and whole-body control chained over 20+ steps and 5+ minutes, such as performing chemistry lab experiments or assembling model airplanes end to end. 2. X-Adapt: adapting from easy data sources (human wearables and simulation) and deploying zero-shot to diverse real robots. 3. Unified Cognition: native integrations with frontier agents for advanced physical reasoning, robust error recovery, and safety monitoring. It's an open-source buffet! We are releasing not only the model weights and X-Adapt recipes (coming December), but also our agentic framework, simulation engine, annotation engine, hardware design, and even training data (100K hours, that's ~11 yrs of experience), all available today! Everything that moves will be autonomous. Anyone can become a roboticist. Lock in, build with us! Deep dive in thread:
194
260
100
1,933
150,337
Ruwan Wickramarachchi retweeted
Introducing REMM: Revisable Embodied Memory for Manipulation. Robot memory needs to do more than retain past observations. If an intervention undoes an earlier step, the robot also needs to work out which completed conditions no longer hold and what needs to be repaired. REMM links task state to supporting visual evidence. When that evidence changes, it revises the affected conclusions and schedules a repair around a frozen G0.5 manipulation policy. Real robot experiments are done on the @GalaxeaDynamics R1Lite. ๐Ÿ“„ arXiv: arxiv.org/abs/2610.12710 ๐ŸŒ Project: revanthsenthil.github.io/remโ€ฆ 1/8 ๐Ÿงต #Robotics #RobotLearning
2
2
3
8
1,129
Excited to share two papers accepted to CoRL 2026! ๐ŸŽ‰ 1. ๐—š๐—ฎ๐—ฃ: ๐—” ๐—š๐—ฟ๐—ฎ๐—ฝ๐—ต-๐—ฎ๐˜€-๐—ฃ๐—ผ๐—น๐—ถ๐—ฐ๐˜† ๐— ๐˜‚๐—น๐˜๐—ถ-๐—”๐—ด๐—ฒ๐—ป๐˜ ๐—ฆ๐—ฒ๐—น๐—ณ-๐—Ÿ๐—ฒ๐—ฎ๐—ฟ๐—ป๐—ถ๐—ป๐—ด ๐—›๐—ฎ๐—ฟ๐—ป๐—ฒ๐˜€๐˜€ ๐—ณ๐—ผ๐—ฟ ๐—ฉ๐—ฎ๐—ฟ๐—ถ๐—ฎ๐˜๐—ถ๐—ผ๐—ป๐—ฎ๐—น ๐—”๐˜‚๐˜๐—ผ๐—บ๐—ฎ๐˜๐—ถ๐—ผ๐—ป ๐—ง๐—ฎ๐˜€๐—ธ๐˜€ Website : graph-robots.github.io/gap/
2
5
108
2. ๐—ฅ๐—ฒ๐—ฐ๐—ผ๐˜ƒ๐—ฒ๐—ฟ, ๐——๐—ถ๐˜€๐—ฐ๐—ผ๐˜ƒ๐—ฒ๐—ฟ, ๐—ฃ๐—น๐—ฎ๐—ป: ๐—Ÿ๐—ฒ๐—ฎ๐—ฟ๐—ป๐—ถ๐—ป๐—ด ๐—ฆ๐—ธ๐—ถ๐—น๐—น๐˜€ ๐—ฎ๐—ป๐—ฑ ๐—–๐—ผ๐—ป๐—ฐ๐—ฒ๐—ฝ๐˜๐˜€ ๐—ณ๐—ฟ๐—ผ๐—บ ๐—ฅ๐—ผ๐—ฏ๐—ผ๐˜ ๐—™๐—ฎ๐—ถ๐—น๐˜‚๐—ฟ๐—ฒ๐˜€ Website: jaraxxus-me.github.io/ReSYNCโ€ฆ
2
40
Announcing Best Paper Award at IROS ReS AI workshop. Best Paper: โ€œiFlax: Neuro-Symbolic Learning for Long-Horizon Task Planning Under Complex Logical Constraintsโ€ Thanks Carnegie Bosch Institute for sponsoring the best paper award. Congratulations! ๐ŸŽ‰ #IROS2026 #ReSAI2026
1
3
4
256
Best Paper Runner-up: โ€œLearning the Right Abstraction for Norm Discoveryโ€ Congratulations to all authors! ๐ŸŽ‰
1
1
87
Interesting panel discussion happening now at IROS ReS AI on what semantics must the data capture for efficient robot learning, and how important is human input? Panelists: @smash0190, @prof_yz , @maojiayuan, and Alessandra Sciutti, moderated by @jeanoh.
1
2
141
Namai Chandra is presenting his work on โ€œActionGround: Training-Free Runtime Refinement of Frozen VLA Policiesโ€ in the second oral session of IROS ReS AI 2026. #ReSAI2026 #IROS2026
1
163
IROS ReS AI second oral talks session just stated with Guy Azran presenting his work on โ€œTask-Aware Conformal State Estimation for Reliable Robotic Planningโ€ #IROS2026 #ReSAI2026
1
1
153
IROS ReS AI Poster session is happening now at Room 411. #IROS2026 #ReSAI2026
1
1
2
199
The fifth keynote at IROS ReS AI 2026 by Dr. Joyce Chai from @michigan_AI on โ€œLanguage Guided Failure Recovery in Robotsโ€ Stay tuned for an exciting poster session, followed by the second oral talks session and panel discussion. #IROS2026 #ReSAI2026
3
4
450
Fourth keynote at the ReS AI workshop by Dr. Yezhou (YZ) Yang (@prof_yz) on โ€œEvaluating AI (and AI Evaluation) as a Science: Towards Science-guided Reliable Agentsโ€. #IROS2026 #ReSAI2026
1
132
First oral talks session at ReS AI 2026. #IROS2026 #ReSAI2026
1
110
Third keynote at the IROS ReS AI workshop by Dr. Jiayuan Mao (@maojiayuan) from the University of Pennsylvania on โ€œIntegrated Learning and Planning with Neuro-Symbolic Conceptsโ€. #ReSAI2026 #IROS2026
214
Second keynote at the IROS ReS AI workshop by Dr. Sebastian Scherer (@smash0190) from the Robotics Institute at CMU on โ€œWhy robots need symbols: finite memory, long horizons, and the limits of scaleโ€ #ReSAI2026
1
1
128
First keynote at IROS ReS AI 2026 workshop by Dr. Alessandra Sciutti from the Italian Institute of Technology on โ€œGrounding Meaning through Embodied and Social Cognitive Developmentโ€ #IROS2026 #ReSAI2026
150
IROS 2026 the first workshop on Embodied Neuro-Symbolic AI for Reliable and Safe Robotics (ReS AI) is happening now in room 411 at the David L. Lawrence Convention Center, in vibrant Pittsburgh! #IROS2026 #ReSAI2026 #EmbodiedNeurosymbolicAI
134