Danya Yao

Papers

Total Citations

H-Index

About

Danya Yao’s research lies at the intersection of computer vision and multi-agent reinforcement learning, with a focus on building robust, perception-driven autonomous systems. In their highly cited 2021 work on self-supervised 3D reconstruction and ego-motion estimation, Yao tackled a fundamental challenge in automated driving and robot navigation: recovering three-dimensional structure from a single monocular camera. By leveraging on-board video, they advanced self-supervised depth estimation methods, addressing issues like occlusions and enabling safer traffic assessment without costly labeled data. This work has garnered 20 citations and remains a touchstone for researchers in geometric deep learning. More recently, Yao has pioneered fault tolerance in multi-agent reinforcement learning (MARL), a critical area for deploying reliable swarms of drones or autonomous vehicles. Their 2025 paper confronts the chaotic state spaces introduced by agent faults, proposing algorithms that help surviving agents extract vital information and adapt. Though still emerging with 3 citations, this work signals a shift toward resilient, real-world AI systems. Yao’s contributions—spanning perception to decision-making—demonstrate a commitment to bridging theory and practice, making their research essential reading for students and engineers building the next generation of autonomous technologies.