Ziyin Xiong  |  熊梓因

I am a second-year MS Research student at the Robotics Institute, Carnegie Mellon University, advised by Prof. Katerina Fragkiadaki. I received my Bachelor's degree in Artificial Intelligence from Tong Class at Yuanpei College, Peking University (2025), with a minor in Economics from the National School of Development.

At Peking University, I was advised by Prof. Yixin Zhu. I also collaborated with Dr. Siyuan Huang and Dr. Tengyu Liu at BIGAI, and Prof. Masayoshi Tomizuka at UC Berkeley.

My goal is to build robots that are useful in everyday life, acting with dexterity, efficiency, and reliability in complex, changing environments.

Email  /  Google Scholar  /  Github  /  X  /  LinkedIn

Research Interest

Large-scale data has driven remarkable progress in robot learning. My research asks how robots could make the most of it: physical understanding that keeps fine-grained actions robust, and reasoning that transfers to new tasks. I pursue this along two directions:

Physically grounded multimodal representations. Effective dexterous manipulation requires understanding scene geometry and the dynamics of physical contact. I study how visual, geometric, and tactile information can be coherently integrated to connect perception, prediction, and action, enabling robots to anticipate and control their interactions with the world.

Scaling reasoning and learning through interaction. I aim to scale robots’ ability to solve novel tasks, reducing reliance on exhaustive demonstration coverage. I'm interested in how foundation-model reasoning can guide planning and exploration, while reinforcement learning enables improvement through experience. Simulation supports scalable interaction, complemented by real-world feedback for adaptation and transfer.

I'm applying for PhD programs starting in Fall 2027 and would be glad to connect about potential opportunities.

News

[Sep 2026] Our work "GeomVLA: Unifying Scene, Motion, and Action in 3D" has been accepted to CoRL 2026.
[July 2026] Our work "Decomposing VAM Uncertainty: Training-Free and Supervised Probes for Robot Failure Detection" has been accepted to the Rethinking Safety Workshop, and selected for an oral presentation at the Trustworthy Embodied Foundation Models Workshop at RSS 2026.
[Oct 2025] Delivered a talk at IROS 2025 on "Ag2x2: Robust Agent-Agnostic Visual Representations for Zero-Shot Bimanual Manipulation."
[Aug 2025] Started my journey at Carnegie Mellon University.
[July 2025] Graduated from Peking University.

We present GeomVLA, a vision-language-action model that unifies scene perception, latent future scene-motion prediction, and action generation in a shared metric 3D frame, which achieves state-of-the-art results on CALVIN and strong performance on LIBERO, RoboTwin2.0, and real-world manipulation without robot-action pretraining.

We present Ag2x2, a framework that advances the autonomous acquisition of bimanual manipulation skills through agent-agnostic and coordination-aware visual representations that jointly encode object and hand motion patterns.

In nature, quadruped animals achieve high-speed locomotion with stable, inertial, and energy-efficient postures. Inspired by these principles, this work aimed to enhance running velocities while ensuring stable robot locomotion by leveraging motion dynamics such as limb movement diagonal symmetry and time-reversal symmetry. Built on Legged Gym, my implementation contributed to a significant improvement in maximum tracking velocity, surpassing the reported state-of-the-art speed of the Go1 robot in simulation.

An online college service system developed by Yuanpei College students, offering a convenient platform for activity room reservations, interest groups, courses, and library services. I am responsible for migrating and integrating the standalone activity room reservation system with the college YPPF website and enhancing the system’s messaging functionality, working with Python.

Services

Conference Reviewer: CoRL 2026.
Workshop Reviewer: Trustworthy Embodied Foundation Models@RSS 2026; CompLearn@ICML 2026.

Miscellaneous

I believe sound is the medium closest to the soul. Honest and meaningful communication can bring the world closer together.

I’m enthusiastic about podcasts for their efficiency and intimate way of sharing knowledge and ideas. I'm preparing my personal podcast channel, exploring topics including geopolitical history, traditional customs and popular culture. Contact me if you're also passionate about these subjects or interested in starting a podcast!