Multi-Agent Reinforcement Learning Enables Cooperative Perception and Monitoring in Dynamic Indoor Environments
A latest arXiv study proposes a cooperative informative sensing framework based on multi-agent reinforcement learning, e…
📂 Research
Latest articles in Research
Comprehensive Survey on VLA Robot Data Infrastructure Released
A new arXiv survey systematically reviews datasets, benchmarks, and data engines for Vision-Language-Action models in ro…
Quantum Computing Threatens Blockchain Security? Latest Research Reveals Challenges and Countermeasures
As quantum computing technology advances rapidly, the cryptographic foundations that blockchain relies upon are facing u…
Novel Lightweight Adhesion Mechanism Enables Drone Landing on Tilted Platforms
A research team has proposed a lightweight toggleable adhesion mechanism that uses motor-driven corkscrew hooks paired w…
RINSE: Improving Imitation Learning Data Quality with Smoothness-Based Metrics
Researchers propose RINSE, a method that automatically evaluates the quality of demonstration data in imitation learning…
RecoverFormer: Teaching Humanoid Robots to Autonomously Recover After Falls
A latest arXiv paper proposes the RecoverFormer framework, an end-to-end policy that enables humanoid robots to autonomo…
AP-BMM: Bayesian Model Merging Approximates the Capability-Efficiency Pareto Frontier of LLMs
A research team proposes the AP-BMM method, which efficiently approximates the Pareto optimal set balancing capability a…
New Breakthrough in Unpowered Fixed-Wing Aircraft Airspeed Safety
A latest arXiv paper proposes an airspeed forward-invariance method for fixed-wing aircraft in unpowered conditions. Thr…
CNN Regression Combined with Rotational Invariance Enables Magnetic Field Indoor Positioning
A latest arXiv paper proposes a magnetic field indoor positioning scheme based on convolutional neural network regressio…
Breakthrough in Satellite-Based Edge AI: Low-Precision-Aware NAS Achieves Deployment Alignment
A latest arXiv paper proposes a deployment-aligned low-precision neural architecture search method for satellite-based e…
Necessary and Sufficient Conditions for Universal Approximation in KAN Networks Have Been Proven
A latest arXiv paper rigorously proves the necessary and sufficient conditions for Kolmogorov-Arnold Networks (KAN) to a…
GS-Playground: A High-Throughput Photorealistic Simulator Empowering Visual Robot Learning
A research team has introduced GS-Playground, a high-throughput photorealistic simulator built on 3D Gaussian Splatting,…