Haoran Yin

Papers

Total Citations

H-Index

About

Haoran Yin is at the forefront of integrating generative AI with autonomous systems, with a primary focus on end-to-end autonomous driving and robotic policy learning. His most-cited work, "DiffusionDrive: Truncated Diffusion Model for End-to-End Autonomous Driving" (2025, 40 citations), introduces a novel approach that harnesses diffusion models—typically used for image generation—to model multi-modal action distributions in driving policies. By proposing a truncated denoising process, Yin significantly reduces computational latency while preserving the model's ability to capture diverse, human-like driving behaviors. This breakthrough addresses a critical bottleneck in deploying diffusion-based policies in real-time autonomous systems. His research demonstrates how generative modeling can bridge the gap between simulation and real-world driving, offering a scalable path toward safer, more adaptive self-driving technology. With his work already gaining rapid attention in the robotics and autonomous driving communities, Yin is establishing himself as a key innovator in the intersection of generative AI and embodied intelligence.