Action Agent pairs LLM-driven video generation with a flow-constrained diffusion transformer to produce velocity commands, raising video success to 86% and delivering 64.7% real-world navigation on a Unitree G1 humanoid.
Title resolution pending
8 Pith papers cite this work. Polarity classification is still indexing.
8
Pith papers citing it
citation-role summary
background 1
citation-polarity summary
years
2026 8verdicts
UNVERDICTED 8roles
background 1polarities
background 1representative citing papers
A variational Gaussian-mixture belief with Gumbel-Softmax and reparameterized sampling enables direct gradient optimization of tail-risk grasping objectives, improving success rates and cutting planning time versus particle-filter baselines.
A belief-space MPPI controller with CVaR safety constraints achieves 82% success and zero contact violations in uncertain robotic insertion simulations, outperforming risk-neutral and chance-constrained baselines.
A CMDPST framework combined with LTLf enables synthesis of resource-aware robust strategies for robots facing both probabilistic and nondeterministic uncertainty.
Derives closed-form solutions for KL-divergence belief merging and a visit-weighted variant, reducing complexity to O(N|S|) and outperforming standard methods in simulations with noisy sensors and long communication gaps.
A voxel-based geometry reconstruction combined with RL-prioritized search detects navmesh inconsistencies with reduced exploration effort while preserving defect coverage.
MORN augments frozen VLM-based object navigation agents with a System 2 meta-controller using Potentiality Index, Persistence Gating, and Evidence Accumulation to improve goal completion rate from 0.23 to 0.30 and reduce wasted steps on the HM3D dataset.
SNGR selectively applies nested sampling to refine iSAM2 posteriors in high-condition-number regions of ambiguous SLAM graphs, yielding better local likelihoods at lower cost than exhaustive non-Gaussian methods.
citing papers explorer
-
Action Agent: Agentic Video Generation Meets Flow-Constrained Diffusion
Action Agent pairs LLM-driven video generation with a flow-constrained diffusion transformer to produce velocity commands, raising video success to 86% and delivering 64.7% real-world navigation on a Unitree G1 humanoid.
-
Variational Neural Belief Parameterizations for Robust Dexterous Grasping under Multimodal Uncertainty
A variational Gaussian-mixture belief with Gumbel-Softmax and reparameterized sampling enables direct gradient optimization of tail-risk grasping objectives, improving success rates and cutting planning time versus particle-filter baselines.
-
Risk-Constrained Belief-Space Optimization for Safe Control under Latent Uncertainty
A belief-space MPPI controller with CVaR safety constraints achieves 82% success and zero contact violations in uncertain robotic insertion simulations, outperforming risk-neutral and chance-constrained baselines.
-
Resource-Constrained Robotic Planning in the face of Mixed Uncertainty
A CMDPST framework combined with LTLf enables synthesis of resource-aware robust strategies for robots facing both probabilistic and nondeterministic uncertainty.
-
Robust Multi-Agent Target Tracking in Intermittent Communication Environments via Analytical Belief Merging
Derives closed-form solutions for KL-divergence belief merging and a visit-weighted variant, reducing complexity to O(N|S|) and outperforming standard methods in simulations with noisy sensors and long communication gaps.
-
Validating Navmesh using Geometry: Voxel-Based Analysis with Prioritized Exploration
A voxel-based geometry reconstruction combined with RL-prioritized search detects navmesh inconsistencies with reduced exploration effort while preserving defect coverage.
-
MORN: Metacognitive Object-Goal Regulation for Resource-Rational Long-Horizon Navigation
MORN augments frozen VLM-based object navigation agents with a System 2 meta-controller using Potentiality Index, Persistence Gating, and Evidence Accumulation to improve goal completion rate from 0.23 to 0.30 and reduce wasted steps on the HM3D dataset.
-
SNGR: Selective Non-Gaussian Refinement for Ambiguous SLAM Factor Graphs
SNGR selectively applies nested sampling to refine iSAM2 posteriors in high-condition-number regions of ambiguous SLAM graphs, yielding better local likelihoods at lower cost than exhaustive non-Gaussian methods.