Shutong JIN's homepage | KTH Royal Institute of Technology

Shutong JIN

Hi! I am a fourth-year PhD student in robotic manipulation at KTH Royal Institute of Technology, advised by Prof. Florian Pokorny and Prof. Erik Elmroth. I was a visiting PhD student at UC Berkeley's AUTOLab, advised by Prof. Ken Goldberg. Previously, I completed my Master’s degree in Computer Control and Automation at Nanyang Technological University. I received my Bachelor’s degree in Electrical and Electronics Engineering from Wuhan University, with a double degree in Robotics from École Centrale. My research is funded by the Wallenberg AI, Autonomous Systems and Software Program (WASP).

[Contact] [Github] [Google Scholar]

Research

My research spans three threads focused on compositional data curation:

Understanding how dataset composition and properties influence robot policies performance through real-world case studies at scale.
Curating additional data to mitigate observed policy performance drops via targeted real-world collection and generative modeling.
Scaling data curation through parallelization in the cloud.

Research Keywords: Robotic Manipulation, Compositional Data Curation, Cloud Robotics.

Academic Service

Teaching: DD2424 Deep Learning in Data Science, DD1420 Foundations of Machine Learning.
Workshop Organization: RoDGE, IROS 2025 Workshop; CloudGripper, CoRL 2026 Demo; Connected-Robots, ICRA 2026 Workshop.
Reviewing Activities: RSS, CoRL, ICRA, IROS, CASE, RA-L, CVPR, ICCV ECCV, WACV.
Supervised Master's Students: Axel Kaliff, Ben Temming, Jin Yang, Pranish Kumar Munnangi.

Publications

	AGRO-SUVIDE: Agentic Robotics for Surgical Viscoelastic Debridement [Link] Shutong Jin, Ziyang Chen, Preethi Satish, Meadow Shen, Gary Guthart, Florian T. Pokorny and Ken Goldberg Preprint
	MACAW: Reliable And Efficient Surgical Debridement Using Monocular Adaptive Compact Attention Windows [Link] Ziyang Chen, Shutong Jin, Preethi Satish, Sareena Mann, Cael Magner, Danyal Fer, Omid Mohareri, Gary Guthart and Ken Goldberg Preprint
	Adversarial Vulnerabilities of Learned Telesurgery Policies [Link] Shutong Jin, Ziyang Chen, Preethi Satish, Paavan Gupta, Florian T. Pokorny and Ken Goldberg Preprint
	RoboLight: A Dataset with Linearly Composable Illumination for Robotic Manipulation [Link] Shutong Jin, Jin Yang, Muhammad Zahid and Florian T. Pokorny 2026 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS).
	D²OFL: Distillation-Enhanced Diffusion Model Assisted One-Shot Federated Learning Obaidullah Zaland, Shutong Jin, Florian T. Pokorny and Monowar Bhuyan IEEE Transactions on Multimedia 2026.
	Physically-based Lighting Generation for Robotic Manipulation [Link] Shutong Jin, Lezhong Wang, Ben Temming and Florian T. Pokorny 2026 IEEE International Conference on Robotics & Automation (ICRA)
	R900: An Analysis of the Cost-Effectiveness of Random Exploration based on 900 Hours of Robotic Stacking [Link] Shutong Jin, Axel Kaliff, Ruiyu Wang, Muhammad Zahid and Florian T. Pokorny Preprint
	Speculative Policy Orchestration: A Latency-Resilient Framework for Cloud-Robotic Manipulation Chanh Nguyen, Shutong Jin, Florian T. Pokorny and Erik Elmroth Preprint
	ReLumix: Extending Image Relighting to Video via Video Diffusion Models [Link] Lezhong Wang, Shutong Jin, Ruiqi Cui, Anders Bjorholm Dahl, Jeppe Revall Frisvad and Siavash Bigdeli Preprint
	One-Shot Federated Learning with Classifier-Free Diffusion Models [Link] Obaidullah Zaland, Shutong Jin, Florian T. Pokorny and Monowar Bhuyan IEEE International Conference on Multimedia & Expo (ICME) 2025
	PACA: Perspective-Aware Cross-Attention Representation for Zero-Shot Scene Rearrangement [Link] Shutong Jin, Ruiyu Wang, Kuangyi Chen and Florian T. Pokorny Proceedings of the IEEE/CVF Winter Conference on Applications of Computer Vision (WACV), 2025
	Feature Extractor or Decision Maker: Rethinking the Role of Visual Encoders in Visuomotor Policies [Link] Ruiyu Wang, Zheyu Zhuang, Shutong Jin, Nils Ingelhag, Danica Kragic and Florian T. Pokorny 2025 IEEE International Conference on Robotics & Automation (ICRA).
	CloudGripper-Push-1K: Understanding the Generalization Gap of Physics and Background Attributes for Robotic Manipulation Shutong Jin, Ruiyu Wang, Muhammad Zahid and Florian T. Pokorny IEEE/RSJ IROS 2024 Workshop on Collecting, Managing, and Utilizing Data through Embodied Robots. Best Poster Award
	CloudGripper-AutoGrasper: A Cloud Robotics Toolkit for Automatic Data Collection Axel Kaliff, Shutong Jin, Muhammad Zahid and Florian T. Pokorny IEEE/RSJ IROS 2024 Workshop on Collecting, Managing, and Utilizing Data through Embodied Robots.
	RealCraft: Attention Control as A Tool for Zero-shot Consistent Video Editing [Link] Shutong Jin, Ruiyu Wang and Florian T. Pokorny 32nd International Conference on Neural Information Processing.
	How Physics and Background Attributes Impact Video Transformers in Robotic Manipulation: A Case Study on Planar Pushing [Link] Shutong Jin, Ruiyu Wang, Muhammad Zahid and Florian T. Pokorny 2024 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS).
	SectionKey: 3-D Semantic Point Cloud Descriptor for Place Recognition [Link] Shutong Jin, Zhenyu Wu, Chunyang Zhao, Jun Zhang, Guohao Peng and Danwei Wang 2022 IEEE/RSJ International Conference on Intelligent Robots and Systems (IROS).