Xiao-Xin Deng

About

Xiao-Xin Deng is a rising researcher at the intersection of robotics, computer vision, and natural language processing, with a primary focus on instruction-driven robotic task planning. His most notable contribution is the development of **GRID (Scene-Graph-based Instruction-driven Robotic Task Planning)**, a pioneering framework that leverages scene graphs—structured representations of objects and their relationships—to enhance how Large Language Models (LLMs) ground complex instructions in real-world environments. Unlike prior methods that rely on raw images, GRID enables robots to reason more effectively about spatial and semantic contexts, significantly improving task execution accuracy. This work, published in 2023, has already garnered early citations, signaling its growing influence in the field. Deng’s research addresses a critical bottleneck in embodied AI: bridging high-level language commands with low-level robotic actions. By integrating scene graphs with LLMs, he offers a more robust and interpretable path toward generalizable robot autonomy. His contributions are particularly relevant for applications in service robotics, manufacturing, and human-robot collaboration, where precise understanding of environment and intent is paramount.

Research Focus

Key Achievements

Top Papers

1. 1
   GRID: Scene-Graph-based Instruction-driven Robotic Task Planning

   2 citations · 2023

Key Collaborators

Contact & Links