Xin-Yi Li

It is argued that harnessing physical-digital AI agents, through persistent coupling of digital intelligence with physical sensing, dynamics, action, and feedback, provides a system-level pathway toward adaptive and continuously evolving UAV autonomy.

Transforming a 6-DoF pose estimation problem into a 4-DoF challenge leads to significant efficiency gains and improved accuracy in robotic applications.

A multimodal diffusion policy outperforms traditional methods, achieving a 78% success rate in industrial cable manipulation tasks, demonstrating a leap in automation capabilities.

Parallel multi-anchor retrieval can dramatically improve the accuracy of legal liability determinations in traffic law by overcoming the limitations of traditional single-axis approaches.

Achieving 97% global executability in robotic assembly, this framework seamlessly integrates cognitive decision-making with physical execution, overcoming traditional limitations.

In high-stakes health contexts, stakeholders demand that trustworthiness in AI systems be inspectable, not just asserted, reshaping how we design health information tools.

HeterSEED achieves state-of-the-art performance on heterophilic heterogeneous graphs by decoupling semantic and structural information, offering a more robust approach than relying on feature similarity alone.

A single system can now generate detection rules across diverse contexts and languages, outperforming pure LLM generation by leveraging dual semantic projection spaces.