Nikhil Madaan
Papers
3
Total Citations
67
H-Index
2
About
Nikhil Madaan is an emerging researcher at the intersection of natural language processing, computer vision, and robotics, with a focused specialization in 3D scene understanding and language-grounded perception for embodied AI systems. His most recognized contribution, **LLM-Grounder**, has garnered over 60 citations and introduces a novel framework that leverages large language models as intelligent agents to perform open-vocabulary 3D visual grounding — enabling robots to locate objects in complex environments based on natural language queries without relying on extensive labeled training data. This work addresses a critical bottleneck in developing household robots capable of navigation, object manipulation, and spatial reasoning. Building on this foundation, Madaan contributed to **3D-GRAND**, a million-scale dataset designed to enhance 3D-language model alignment while reducing hallucination — a persistent challenge when adapting LLMs to physical, three-dimensional environments. Together, these works position Madaan as a contributor to the growing field of 3D-LLMs and embodied intelligence. His research is particularly valuable for students exploring how large language models can be grounded in physical reality, bridging the gap between abstract language understanding and real-world robotic perception.
Research Focus
Key Achievements
Top Papers
- 1
- 2
- 3