Tsinghua AI

Agentic robots are no longer just predicting actions, but actively orchestrating their own learning and execution through goal-directed reasoning.

Imagined failures in a world model can pinpoint exactly which robot skill components need new demonstrations, boosting success from 13% to 75% with minimal data.

Robots no longer need human intervention to learn from their mistakes: foundation-model-guided exploration over modular skills can autonomously synthesize corrective rollouts that distill directly into low-level policy weights.

Purely autoregressive VLAs no longer require awkward hybrid diffusion heads to achieve fine-grained robot control: structuring discrete action tokenization as causally annealed flow matching recovers continuous diffusion precision while preserving end-to-end token efficiency.

The concept of traversability is extended from robot-centric to human-centric by accounting for the operator's cognitive states, such as attention, workload, and risk tolerance or awareness, termed tele-traversability, and is revisited and extended to teleoperation settings.

A synthetic-to-real framework coupling physically validated data generation with a hierarchical, diffusion-based controller is proposed, which outperforms state-of-the-art manipulation policies in transport smoothness and dynamic stability.

Uncertainty-Guided Refinement (UGR), a sparse refinement framework for chunk-based visuomotor policies, which first predicts a full action chunk, estimates per-step temporal uncertainty from the coarse hidden states, and applies residual correction only to the most uncertain timesteps selected by a binary mask.

TANGO enables humanoid robots to navigate cluttered spaces with unprecedented language-guided precision, achieving state-of-the-art performance without real-world training.

This work presents the first synergistic framework that tightly couples hierarchical memory with proactive tool invocation in a closed reasoning loop, and reveals the complementary roles of hierarchical memory and tool strategies.

Robots can now perform complex manipulation tasks more effectively by understanding and responding to contact forces without additional sensors.

CAER redefines how we train world models by ensuring that only the most relevant tokens drive learning, leading to substantial gains in video generation quality.

A novel congestion-aware scheduling method reduces delivery times by over 16% in semiconductor fabs without sacrificing throughput.

A closed-loop world model that learns from both successes and failures could revolutionize dexterous manipulation in robotics.

Robots can now autonomously evolve their manipulation skills in real-time, leveraging past interactions without needing to update their underlying policy models.

Recalibrating supervision in world model training can drastically enhance interaction fidelity and physical plausibility without the need for external representations.

EMERGE-Policy reveals that a robot's intelligence can thrive through a collaborative network of specialized agents rather than a monolithic policy.

Achieving minute-long real-time interactive world generation while maintaining geometric consistency and dynamic subject identity is now a reality with Matrix-Game 3.5.

Riemann-1.0 transforms embodied intelligence by unifying robot policy execution and world simulation, achieving unprecedented success rates in real-world manipulation tasks.

LAC enables humanoid robots to dynamically adapt their compliance to external forces, significantly improving their interaction capabilities in real-world tasks.

Agents using ParallelWorld can efficiently evaluate multiple future trajectories, leading to superior decision-making in complex environments.

Typed spatial readouts can elevate robotic execution success rates by over 50% while significantly reducing processing time.

Command-inconsistent memory can lead to a 15.6% increase in collision rates during autonomous driving, but MomADv2 effectively filters this noise for safer long-horizon planning.

Achieving a 99.8% success rate in reconnecting initially disconnected convex regions transforms the feasibility of GCS-based motion planning.

DECOWAM achieves a 21.7% reduction in action prediction error while maintaining robust task performance, showcasing the power of embodiment-aware factorization in mobile manipulation.

Policies may succeed in tasks but still violate deformation tolerances, revealing a critical gap in current evaluation methods for deformable-object manipulation.

Separating spatial and temporal modeling in action recognition leads to significant performance gains, challenging the effectiveness of existing implicit coupling methods.

HODAgent's unified architecture enables humanoid robots to adaptively manage service tasks in real-time, achieving up to 92% success rates on physical platforms.

Selectively routed stereo evidence boosts humanoid VLA control success rates, achieving 100% grasp success even under severe occlusion.

A non-invasive framework that enhances pretrained generative policies with force control, achieving remarkable improvements in contact stability and execution speed.

Guiding dexterous grasp generation with arm-aware constraints boosts feasibility in complex environments, outperforming traditional methods reliant on hand-centric models.

Closed-loop learning in embodied agents can lead to a staggering 11.1x speedup in inference while achieving record performance on complex tasks.

SparkVLA redefines task execution in hierarchical VLA systems, achieving over 30% improvement in success rates by integrating stop and action length decisions into a single ranking process.

Vision-based tactile sensors could revolutionize robotic interaction by providing high-resolution tactile data that enhances perception and manipulation capabilities.

HumanScore reveals that traditional kinematic metrics overlook critical failures in humanoid motion tracking, such as unstable support and incorrect contacts.

Robots can now predict not just immediate actions but also the next stages of complex tasks, leading to more efficient manipulation strategies.

JEPA-WAM achieves a remarkable 79.2% on LIBERO-Plus without large-scale pretraining, setting a new benchmark for efficient robot control.

SLIM achieves state-of-the-art performance in robot manipulation with just 0.5B parameters, outperforming larger models while slashing GPU memory usage and inference latency.

OnEvoMemory allows robots to evolve their memory in real-time, leading to improved task performance and reduced redundancy in long-horizon manipulation tasks.

SkillMemo transforms robotic manipulation by enabling models to leverage reusable skill structures, leading to unprecedented compositional generalization in complex tasks.

GeniWorld achieves robust zero-shot generalization in robotic manipulation, outperforming traditional models even with minimal training data.

Trajectory scoring in aerial navigation can be revolutionized by focusing on unexplainable prediction discrepancies, leading to more robust and efficient UAV navigation.

RTCF boosts the success of frozen VLA policies by leveraging past experiences without the need for retraining or extra GPU power.