Dongyu Ji

Papers

Total Citations

H-Index

About

Dongyu Ji is an emerging researcher specializing in embodied AI, robot navigation, and vision-and-language understanding. His work sits at the intersection of computer vision, natural language processing, and autonomous robotics, with a particular focus on enabling intelligent agents to navigate complex real-world environments using both visual perception and linguistic guidance. Ji's most notable contribution, "Weakly-Supervised Multi-Granularity Map Learning for Vision-and-Language Navigation" (2022, 16 citations), addresses the challenge of training robot agents to follow natural language instructions in dynamic environments. By developing multi-granularity map representations, his approach significantly advances the accuracy and efficiency of instruction-following navigation systems. Complementing this, his work on "Learning Active Camera for Multi-Object Navigation" (2022, 8 citations) tackles the underexplored problem of adaptive camera control, demonstrating that active sensing strategies can dramatically improve a robot's ability to locate multiple objects autonomously. Together, these contributions reflect Ji's broader mission to bridge the gap between language understanding and physical navigation in robotics. Though early in his research career, his work has already garnered meaningful attention from the community, establishing him as a promising voice in the rapidly growing field of embodied intelligence and autonomous agent research.