Foresight detects failures in long-horizon robotic manipulation using latents from action-conditioned world models trained only on task-level labels and calibrated via functional conformal prediction.
S ¸ucan, Mark Moll, and Lydia E
7 Pith papers cite this work. Polarity classification is still indexing.
7
Pith papers citing it
representative citing papers
A hierarchical multi-robot motion planner that refines workspace decompositions to enable scalable coordination through discrete search over smaller decoupled subproblems.
TwinRL expands RL exploration via digital twin reconstruction and twin RL warm-up to guide real-world learning, reaching near-100% success with 20 minutes of on-robot time across four tasks.
ActivePusher integrates residual-physics modeling with uncertainty-based active learning to improve data efficiency and planning success rates for nonprehensile manipulation in simulation and real-world settings.
Neural CDF barriers enable efficient planning and distributionally robust safe control for manipulators in cluttered dynamic environments using only point-cloud observations.
GeneralVLA-2 introduces GeoFuse-MV3D for improved multi-view 3D reconstruction and a governed memory system, demonstrating modest gains on 3D object and task benchmarks.
citing papers explorer
-
Foresight: Failure Detection for Long-Horizon Robotic Manipulation with Action-Conditioned World Model Latents
Foresight detects failures in long-horizon robotic manipulation using latents from action-conditioned world models trained only on task-level labels and calibrated via functional conformal prediction.
-
Scalable Multi-robot Motion Planning via Hierarchical Subproblem Expansion and Workspace Decomposition Refinement
A hierarchical multi-robot motion planner that refines workspace decompositions to enable scalable coordination through discrete search over smaller decoupled subproblems.
-
TwinRL: Digital Twin-Driven Reinforcement Learning for Real-World Robotic Manipulation
TwinRL expands RL exploration via digital twin reconstruction and twin RL warm-up to guide real-world learning, reaching near-100% success with 20 minutes of on-robot time across four tasks.
-
ActivePusher: Active Learning and Planning with Residual Physics for Nonprehensile Manipulation
ActivePusher integrates residual-physics modeling with uncertainty-based active learning to improve data efficiency and planning success rates for nonprehensile manipulation in simulation and real-world settings.
-
Neural Configuration-Space Barriers for Manipulation Planning and Control
Neural CDF barriers enable efficient planning and distributionally robust safe control for manipulators in cluttered dynamic environments using only point-cloud observations.
-
GeneralVLA-2: Geometry-Aware Reconstruction and Governed Memory for Robot Planning
GeneralVLA-2 introduces GeoFuse-MV3D for improved multi-view 3D reconstruction and a governed memory system, demonstrating modest gains on 3D object and task benchmarks.