Xinhang Song

Papers

Total Citations

H-Index

About

Xinhang Song is a researcher whose work bridges computer vision and embodied AI, with a focus on scene understanding and human-robot interaction. His early contributions include participation in the ImageCLEF 2013 Robot Vision Challenge with the MIAR ICT team, where he tackled indoor scene classification and object recognition—foundational problems in visual perception. More recently, Song has advanced the critical area of explainable AI for embodied agents. In his 2023 work, "Generating Explanations for Embodied Action Decision from Visual Observation," he addresses the challenge of building trust between humans and autonomous systems like robots and self-driving cars. By developing methods that produce natural language explanations for visual-based decisions, Song is making embodied agents more transparent and accessible, especially for non-expert users. While his citation counts are still building—with his most cited work at 4 citations—his focus on explainability places him at an important intersection of vision, language, and robotics. As the demand for trustworthy autonomous systems grows, Song’s research into generating clear, human-readable justifications for machine actions is poised to become increasingly influential in shaping how we collaborate with intelligent machines.