Current research in robotics is increasingly focused on enhancing autonomy and adaptability in dynamic environments, with significant strides in zero-shot learning and human-robot collaboration. Recent work on scene reconstruction and robot grasping employs differentiable neuro-graphics to enable robots to interact with previously unseen objects without extensive training data, addressing a critical barrier to deployment in novel settings. Similarly, advancements in human-aware navigation for UAVs demonstrate the ability to autonomously assist in emergency scenarios by generating trajectories directly from visual input, enhancing operational efficiency. Multi-task model-based reinforcement learning is gaining traction, emphasizing the importance of task diversity over sheer data volume to improve learning efficiency and robustness. Additionally, innovative frameworks for humanoid interaction and path planning in agriculture are pushing the boundaries of how robots perceive and interact with their environments. Collectively, these developments signal a shift toward more data-efficient, interpretable, and responsive robotic systems capable of tackling complex real-world challenges.
Top papers
- Learning Whole-Body Human-Humanoid Interaction from Human-Human Demonstrations(8.0)
- Differentiable Inverse Graphics for Zero-shot Scene Reconstruction and Robot Grasping(8.0)
- HumanDiffusion: A Vision-Based Diffusion Trajectory Planner with Human-Conditioned Goals for Search and Rescue UAV(8.0)
- Disentangling perception and reasoning for improving data efficiency in learning cloth manipulation without demonstrations(8.0)
- Scaling Tasks, Not Samples: Mastering Humanoid Control through Multi-Task Model-Based Reinforcement Learning(8.0)
- IROSA: Interactive Robot Skill Adaptation using Natural Language(7.0)
- TOPReward: Token Probabilities as Hidden Zero-Shot Rewards for Robotics(7.0)
- Learning Object-Centric Spatial Reasoning for Sequential Manipulation in Cluttered Environments(7.0)
- Non-Markovian Long-Horizon Robot Manipulation via Keyframe Chaining(7.0)
- VISTA: Enhancing Visual Conditioning via Track-Following Preference Optimization in Vision-Language-Action Models(7.0)
- Optimizing Path Planning using Deep Reinforcement Learning for UGVs in Precision Agriculture(7.0)
- Closed-Loop Action Chunks with Dynamic Corrections for Training-Free Diffusion Policy(7.0)
- STaR: Scalable Task-Conditioned Retrieval for Long-Horizon Multimodal Robot Memory(7.0)
- Accelerating Robotic Reinforcement Learning with Agent Guidance(7.0)
- Low Cost, High Efficiency: LiDAR Place Recognition in Vineyards with Matryoshka Representation Learning(7.0)
- H-AIM: Orchestrating LLMs, PDDL, and Behavior Trees for Hierarchical Multi-Robot Planning(7.0)
- PUMA: Perception-driven Unified Foothold Prior for Mobility Augmented Quadruped Parkour(7.0)
- Hybrid guided variational autoencoder for visual place recognition(6.0)
- Boosting Deep Reinforcement Learning with Semantic Knowledge for Robotic Manipulators(6.0)
- DECO: Decoupled Multimodal Diffusion Transformer for Bimanual Dexterous Manipulation with a Plugin Tactile Adapter(6.0)
- Mean-Flow based One-Step Vision-Language-Action(6.0)
- MeshMimic: Geometry-Aware Humanoid Motion Learning through 3D Scene Reconstruction(6.0)
- Retrieval-Augmented Robots via Retrieve-Reason-Act(6.0)
- Self-Correcting VLA: Online Action Refinement via Sparse World Imagination(6.0)
- Co-jump: Cooperative Jumping with Quadrupedal Robots via Multi-Agent Reinforcement Learning(5.0)
- Cross-Embodiment Offline Reinforcement Learning for Heterogeneous Robot Datasets(5.0)
- Training and Simulation of Quadrupedal Robot in Adaptive Stair Climbing for Indoor Firefighting: An End-to-End Reinforcement Learning Approach(5.0)
- Flow Policy Gradients for Robot Control(5.0)
- GarmentPile++: Affordance-Driven Cluttered Garments Retrieval with Vision-Language Reasoning(5.0)
- Interaction-Aware Whole-Body Control for Compliant Object Transport(5.0)
- Residual RL--MPC for Robust Microrobotic Cell Pushing Under Time-Varying Flow(5.0)
- Lifelong Language-Conditioned Robotic Manipulation Learning(4.0)
- Training slow silicon neurons to control extremely fast robots with spiking reinforcement learning(3.0)
- Learning Transferability: A Two-Stage Reinforcement Learning Approach for Enhancing Quadruped Robots' Performance in U-Shaped Stair Climbing(3.0)
- CMR: Contractive Mapping Embeddings for Robust Humanoid Locomotion on Unstructured Terrains(3.0)
- Self-adapting Robotic Agents through Online Continual Reinforcement Learning with World Model Feedback(2.0)