Yinan Zheng

Papers

Total Citations

H-Index

About

Yinan Zheng is a rising researcher at the forefront of embodied AI and foundation models, with a focus on bridging the gap between internet-scale data and real-world robotic agents. Their most-cited work, “Universal Actions for Enhanced Embodied Foundation Models” (2025, 6 citations), tackles a critical bottleneck in robotics: while large language and vision models thrive on diverse, crowd-sourced data, embodied agents struggle with incompatible action spaces across datasets. Zheng’s key contribution is the introduction of a universal action representation that enables training on heterogeneous embodied data without costly re-annotation, effectively scaling imitation learning across tasks and platforms. This approach promises to unlock more generalist robots capable of adapting to novel environments. Though early in their career, Zheng’s work addresses a foundational challenge in embodied AI—data heterogeneity—and has already garnered attention for its potential to democratize robotic learning. Their research sits at the intersection of computer vision, reinforcement learning, and robotics, offering a practical pathway toward more robust, data-efficient embodied agents.