Xiaojun Chang

Monash University, University of Science and Technology of China

Papers

Total Citations

165

H-Index

About

Xiaojun Chang is a prominent researcher specializing in embodied artificial intelligence, vision-and-language navigation, and deep learning for autonomous agents. His work sits at the exciting intersection of computer vision, natural language processing, and robotics, addressing some of the most ambitious challenges in intelligent system design. Chang's most influential contribution, "SOON: Scenario Oriented Object Navigation with Graph-based Exploration" (2021, 115 citations), tackled a fundamental limitation in visual navigation research by enabling agents to navigate toward language-guided targets from arbitrary starting points within complex 3D environments — mimicking the flexibility of human navigation in a way that fixed-start benchmarks could not capture. His comprehensive survey on deep learning for embodied vision navigation further established him as a synthesizer of this rapidly evolving field, providing students and researchers with a valuable roadmap of the discipline. More recently, his work on NavCoT (2025, 32 citations) demonstrates his forward-looking engagement with large language models, exploring how disentangled reasoning can enhance agent decision-making in complex environments. Collectively, Chang's research has meaningfully advanced the scientific community's understanding of how intelligent agents can perceive, reason, and navigate real-world environments, making his publications essential reading for anyone entering embodied AI research.