Dai Yinlong

"A worshiper of science,
a slave to freedom."

Dai, Yinlong
戴尹龙

Hello! My name is Yǐnlóng DAI. I am a PhD student at Virginia Tech. I am currently working on human robot interaction in the Collaborative Robotics Lab (Collab) directed by Dr. Dylan Losey.

Prior to this, I received my Bachelor's degree with Honors in Computer Science and a minor in Robotics from New York University. I was an undergraduate research assistant at the Generalizable Robotics and AI Lab (GRAIL) under the supervision of Prof. Lerrel Pinto, where I worked on robotic dexterous manipulation. I was also fortunate to work with Prof. Gizem Kayar on surgical simulation and Prof. Qi Lei on medical image classification.

Logs

2026.1 Prepare Before You Act: Learning From Humans to Rearrange Initial States accepted to ICRA 2026.
2024.5 🎓I graduated from New York University.

Research

I'm interested in robotics, computer vision and imitation learning. I work at the intersection of Machine learning and Robotics.

	From Local Corrections to Generalized Skills: Improving Neuro-Symbolic Policies with MEMO Benjamin A. Christie, Yinlong Dai, Mohammad Bararjanianbahnamiri, Simon Stepputtis, Dylan P. Losey Under Review, 2026 code / arXiv By aggregating natural language corrections from multiple users across multiple tasks into a skillbook, we create and refine a continuously evolving set of generalized skills.
	Language Movement Primitives: Grounding Language Models in Robot Motion Yinlong Dai, Benjamin A. Christie, Daniel J. Evans, Dylan P. Losey, Simon Stepputtis Under Review, 2026 code / arXiv We propose Language Movement Primitives (LMP), a framework that bridges the gap between high-level reasoning and low-level robotic control through the use of Dynamic Movement Primitives.
	Prepare Before You Act: Learning From Humans to Rearrange Initial States Yinlong Dai, Andre Keyser, Dylan P. Losey ICRA, 2026 code / arXiv In this work, we present ReSET, a framework that learns to restructure the environment, bringing diverse start states into a familiar and manageable distribution before executing the actual task.
	CIVIL: Causal and Intuitive Visual Imitation Learning Yinlong Dai , Robert Ramirez Sanchez , Ryan Jeronimus, Shahabedin Sagheb, Cara M. Nunez, Heramb Nemlekar, Dylan P. Losey Under Review, 2025 code / arXiv A framework designed to enable real robots to learn from multimodal visual-language instruction data using imitation learning.
	HuDOR: Bridging the Human to Robot Dexterity Gap through Object-Oriented Rewards Irmak Guzey, Yinlong Dai, Georgy Savva, Raunaq Bhirangi, Lerrel Pinto ICRA, 2025 code / arXiv In this work, we present HuDOR, a technique that enables online fine-tuning of policies by directly computing rewards from human videos.
	OPEN TEACH: A Versatile Teleoperation System for Robotic Manipulation Aadhithya Iyer, Zhuoran Peng, Yinlong Dai, Irmak Guzey, Siddhant Haldar, Soumith Chintala, Lerrel Pinto CORL, 2024 code / arXiv A new teleoperation system leveraging VR headsets to immerse users in mixed reality for intuitive robot control.
	See to Touch: Learning Tactile Dexterity through Visual Incentives Irmak Guzey, Yinlong Dai, Ben Evans, Soumith Chintala, Lerrel Pinto ICRA, 2024 code / arXiv In this work, we present Tactile Adaptation from Visual Incentives (TAVI), a framework that enhances tactile-based dexterity by optimizing dexterous policies using vision-based rewards.
	Exploring Minimally Sufficient Representation in Active Learning through Label-Irrelevant Patch Augmentation Zhiyu Xue, Yinlong Dai, Qi Lei Conference on Parsimony and Learning, PMLR, 2024 PDF In this work, we intend to learn a "good" feature representation that is both sufficient and minimal, facilitating effective AL for medical image classification.