This work proposes Fast-TD-MPC, a lightweight framework that adaptively routes between fast policy execution and test-time planning, reserving costly deliberation for states where it is most needed, and delivers competitive task performance across 103 continuous control tasks while achieving up to ~4x faster inference.
FedDRMan, a federated subspace-guided distillation framework for heterogeneous robot manipulation, substantially improves knowledge transfer and consistently outperforms strong federated baselines achieving a peak mean success rate of 80.7%.
Biprodip Pal, Kaushik Roy, Yan-Ming Zhu et al.· 0 citations
Vision-Language-Action (VLA) models have shown strong performance on robotic manipulation, but they often struggle to generalize to unseen tasks, configurations, and long-horizon settings. A key challenge is that VLAs overfit to training scenes and fail to follow novel language instructions. Off-the-shelf vision-langua...
This work proposes a calibration-free procedure for the estimation of contact surface normals using Universal Photometric Stereo neural networks, and demonstrates that with sufficient illumination settings surface normals could be estimated using a model trained solely on synthetic data.
This work introduces manifold-stable flow matching (MSFM), which can start from an arbitrary ambient prior, not necessarily supported on the manifold, and guarantees manifold invariance and transverse convergence to the manifold within a desired time window.
Amirhossein Nazerian, A. Pezeshki, Jian-Guo Zhao· 0 citations
Adjoint Guidance Flow is proposed, which is a deterministic optimal control problem, whose optimal guidance is a costate that carries the terminal critic gradient back through the remaining flow, and regress the guidance network onto this costate while keeping both the VLA and critic frozen.
Jeongsol Kim, Youngjun Jun, Kyumin Choi et al.· 0 citations
This work proposes and evaluates State-Conditioned Shooting (SCOOT), a novel DRL algorithm that builds on advantage-weighted regression (AWR) with three key modifications, and showcases the features’ performance in learning physically-based billiard shots demonstrating high action precision and discovering multiple sho...
N. H. Kim, Markus Kirjonen, Perttu Hämäläinen· Motion in Games· 0 citations
Rate-limited multi-agent teams raise three questions the emergent-communication literature has answered only empirically: what an optimal message should encode, what compression costs over a horizon, and when a learned protocol is unique enough for a teammate to read, then measure how far reinforcement learning falls s...
Online safe reinforcement learning (RL) seeks policies that maximize reward while satisfying safety constraints. A popular line of research in safe RL relaxes safety to a soft expected-cost constraint and solves the resulting Constrained Markov Decision Process via primal-dual Lagrangian updates that only enforce safet...
Bo-Yang Li, Matthew Kim, Sylvia L. Herbert· 0 citations
When the human reference scores zero on a metric, the released GTRS-Dense label generator for NAVSIM marks every candidate trajectory in the scene as passing it, and the consequences for supervision are measured.
Jia-Xuan Guo, Jing-Xin Yang, Jia-Qi Ye et al.· 0 citations
M measuring the cost of a memory cut by training with it from initialization, and comparing backward graphs by the updates the optimizer applies rather than by raw gradients suggest the cost of a memory cut should be measured by training with it from initialization.
Xing-Jian Li, Yi Han, Jian-Hua Z. Huang· 0 citations
World models simulate the consequences of action candidates, but good planning need not preserve every physical distinction required for accurate prediction. We formalize this gap through a hierarchy of mechanism, response, and decision sufficiency. Given a candidate set, the planning query determines which physical va...
Rong-Zhe Wei, Hans Hao-Hsun Hsu, Pei-Zhi Niu et al.· 1 citation
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.
