In robotic manipulation, the tight coupling between grasping and motion planning often obscures the true source of failure, leading to inefficient trial-and-error. To enable efficient long-horizon manipulation, we propose GTP-FA (Grasp-Then-Plan with Failure Attribution), a task-oriented two-stage 'grasp-then-plan' fra...
Jiahao Xu, Peiyuan Wang, Hanzhuo Zhang et al.· 0 citations
Vision-Language-Action (VLA) models remain brittle in long-horizon, contact-rich manipulation because success-only imitation provides little supervision for execution drift, while failed rollouts are often discarded. We introduce RePO-VLA, a recovery-driven policy optimization framework that assigns distinct roles to s...
Weijia Liufu, Xiaoyu Guo, Ruiyi Chen et al.· 0 citations
Planning with world models offers a powerful paradigm for robotic control. Conventional approaches train a model to predict future frames conditioned on current frames and actions, which can then be used for planning. However, the objective of predicting future pixels is often at odds with the actual planning objective...
Jacob Berg, Chuning Zhu, Yanda Bao et al.· 0 citations
Vision-Language-Action (VLA) models have made substantial progress in robotic manipulation, but their generalization remains constrained by the diversity and coverage of action-labeled robot demonstrations. Human videos provide much broader behavioral and environmental experience, while learning from them faces a funda...
Chaehyun Kim, Chi-Pin Huang, Jisoo Kim et al.· 0 citations
Reach audiences
Advertise in front of researchers, engineers, and readers.
Manipulation policies are evaluated predominantly by binary task success, which cannot tell whether the commanded gesture is the one the object's kinematics admit. We introduce Constraint-Referenced Strategy (CRS), which projects the commanded end-effector twist onto the object's joint axis, separating progress along t...
Baptiste Bonin, Caro Strickland, Audrey Durand· 0 citations
Adaptive Residual Correction is introduced, a lightweight post-training module that uses reinforcement learning (RL) to learn a residual correction on top of a frozen policy while using sparse human corrections to structure autonomous exploration.
Anukriti Singh, Maxime Caniot, Pratap Tokekar et al.· 0 citations
Failure diagnosers for robot manipulation are increasingly vision--language models (VLMs), trained and evaluated on single-arm failures produced by perturbing a successful demonstration. Bimanual tool use adds failures of other kinds: a tool that slips out of its socket while still held, two correct gestures made at di...
Baptiste Bonin, Caro Strickland, Audrey Durand· 0 citations
A fixed-wing interceptor is presented that is controlled by a neural network that maps filtered monocular vision observations directly to actuator commands and is able to intercept targets both in simulation and in real flight.
D. V. van Hamersveld, E. Smeur, Iurii Kriachko· 0 citations
Quality-Controllable Policy is introduced, a robot policy that enables inference-time control over both individual quality attributes and their combinations and their combinations, and an iterative hindsight relabeling loop is proposed to extend the controllable quality range beyond initial demonstrations.
Sinjae Kang, Dongyeeb Shin, Jae-Joon Kim et al.· 0 citations
Real-to-Sim-to-Real (Real2Sim2Real) enables structured world representations for planning, interpretability, and safe execution. Most existing approaches rely on passive perception, which creates real2sim gaps due to imperfect observation. Inspired by how humans perceive actively through actions, we propose ActiveAlign...
Ri-Zhao Qiu, Bowen Wen, Yifeng Jiang et al.· 0 citations
Pretrained video generation models encode rich priors about physical interactions, but leveraging them for robot control typically requires either resource-intensive video generation at test time or action heads trained from scratch that do not fully utilize the pretrained backbone. We introduce Tokenizing Actions as V...
Jeremy A. Collins, Seungjae Lee, Rachanon Wachakorn et al.· 0 citations
A new system identifies and corrects unreliable object descriptions stored in a robot’s memory, improving the robot’s ability to interact with its environment.
Robots are getting smarter, but how can their hardware match that growth? New Microsoft Research findings show that moving AI inference beyond the robot can improve task success, boost efficiency, and support more advanced physical AI workloads. The post Offloaded inference for real-world physical AI robotics appeared first on Microsoft Research.
