Anukriti Singh

Papers

Total Citations

H-Index

About

Anukriti Singh is a rising researcher at the intersection of computer vision, robotics, and machine learning, with a focus on enabling more efficient and intuitive robot learning. Her work addresses two critical bottlenecks in autonomous systems: navigation in unknown environments and the data-hungry nature of robotic manipulation. In her 2024 paper, “Pre-Trained Masked Image Model for Mobile Robot Navigation,” Singh pioneered the use of pre-trained vision models to predict structural patterns in environments from partial observations, allowing robots to build and use 2D top-down maps more intelligently. This approach reduces reliance on exhaustive sensor data, making real-time navigation more robust. Her 2025 work, “Sketch-to-Skill,” introduces a paradigm shift in robot training by leveraging simple human-drawn trajectory sketches to bootstrap policy learning. This method dramatically lowers the barrier to entry for Imitation and Reinforcement Learning, cutting the need for expert demonstrations or costly environmental rollouts. With both papers already garnering early citations, Singh’s contributions are poised to democratize robot learning, making it accessible to non-experts while advancing the efficiency of autonomous systems in real-world settings.