Mateo Guaman Castro

I am a PhD student at the University of Washington working on Robotics and Machine Learning, advised by Prof. Byron Boots and Prof. Abhishek Gupta.

Previously, I was a Master's student in Robotics in the Robotics Institute at Carnegie Mellon University, advised by Prof. Sebastian Scherer as a member of the AirLab and the Field Robotics Center.

Before that, I studied electrical engineering at Tufts Univesity, where I recieved my BS. I was a researcher in the Multimodal Learning, Interaction, and Perception Lab (MuLIP) advised by Dr. Jivko Sinapov and Dr. Evana Gizzi. I was also a research intern working with Dr. Howie Choset in the Biorobotics lab at Carnegie Mellon University, and Dr. Guillaume Sartoretti in the Multi-Agent Robotic Motion (MARMoT) Lab at the National University of Singapore.

Email  /  CV (Dec. 2024) /  Google Scholar  /  Twitter  /  Github

profile photo
Research

My research lies at the intersection of Robot Learning, Reinforcement Learning, VLMs, and Field Robotics. I am interested in reinforcement learning and representation learning for fast adaptation and generalization in unstructured environments. My ultimate goal is to develop robots that adapt fast during deployment, probe when uncertain, and make mistakes only once, if at all.

News
  • September 2023: I joined UW as a PhD student in Robotics!
  • July 2023: I attended ICML 2023, where I served as Social Chair at the LatinX in AI workshop.
  • July 2023: I defended my Master's thesis at Carnegie Mellon University! Here is the recorded video.
  • May 2023: I presented two conference papers and one workshop paper at ICRA 2023, my first ever in-person conference.
Agile Continuous Jumping in Discontinuous Terrains
Yuxiang Yang, Guanya Shi, Changyi Lin, Xiangyun Meng, Rosario Scalise, Mateo Guaman Castro, Wenhao Yu, Tingnan Zhang, Ding Zhao, Jie Tan, Byron Boots
Submitted to International Conference on Robotics and Automation (ICRA), 2025
arXiv / pdf / project page

We developed a system that enables quadrupedal robots to perform continuous, precise jumps across challenging terrains like stairs and stepping stones, achieving unprecedented agility.

DROID: A Large-Scale In-the-Wild Robot Manipulation Dataset
DROID Dataset Team
Robotics: Science and Systems (RSS), 2024
arXiv / pdf / project page

We present a large dataset for robot learning.

Open X-Embodiment: Robotic Learning Datasets and RT-X Models
Open X-Embodiment Collaboration
International Conference on Robotics and Automation (ICRA), 2024
IEEE ICRA Best Conference Paper Award
arXiv / pdf / project page

We present a large dataset for robot learning.

TartanDrive 2.0: More Modalities and Better Infrastructure to Further Self-Supervised Learning Research in Off-Road Driving Tasks
Matthew Sivaprakasam, Parv Maheshwari, Mateo Guaman Castro, Samuel Triest, Micah Nye, Steven Willits, Andrew Saba, Wenshan Wang, Sebastian Scherer
International Conference on Robotics and Automation (ICRA), 2024
arXiv / pdf / IEEE / project page / video

We present a dataset for off-road driving with multiple modalities at high speed.

How Does It Feel? Self-Supervised Costmap Learning for Off-Road Vehicle Traversability
Mateo Guaman Castro, Samuel Triest, Wenshan Wang, Jason M. Gregory, Felix Sanchez, John G. Rogers III, Sebastian Scherer
International Conference on Robotics and Automation (ICRA), 2023
arXiv / pdf / IEEE / project page / short video / long video

We propose a self-supervised method to predict traversability costmaps by combining exteroceptive environmental information with proprioceptive terrain interaction feedback.

Learning Risk-Aware Costmaps via Inverse Reinforcement Learning for Off-Road Navigation
Samuel Triest, Mateo Guaman Castro, Parv Maheshwari, Matthew Sivaprakasam, Wenshan Wang, Sebastian Scherer
International Conference on Robotics and Automation (ICRA), 2023
arXiv / pdf / IEEE

We present an inverse reinforcement learning-based method that efficiently predicts uncertainty-aware costmaps for off-road traversability via conditional value-at-risk (CVaR).

TartanDrive 1.5: Improving Large Multimodal Robotics Dataset Collection and Distribution
Matthew Sivaprakasam, Samuel Triest, Mateo Guaman Castro, Micah Nye, Mukhtar Maulimov, Cherie Ho, Parv Maheshwari, Wenshan Wang, Sebastian Scherer
Workshop on Pretraining for Robotics (PT4R), International Conference on Robotics and Automation (ICRA), 2023
pdf

In this work we discuss the improvements to our previous dataset, TartanDrive.

Toward Life-Long Creative Problem Solving: Using World Models for Increased Performance in Novelty Resolution
Evana Gizzi, Wo Wei Lin, Mateo Guaman Castro, Ethan Harvey, Jivko Sinapov
International Conference on Computational Creativity (ICCC), 2022
pdf

In this work, we investigate methods for life-long creative problem solving (LLCPS), with the goal of increasing CPS capability over time.

A Novelty-Centric Agent Architecture for Changing Worlds
Faizan Muhammad, Vasanth Sarathy, Gyan Tatiya, Shivam Goel, Saurav Gyawali, Mateo Guaman Castro, Jivko Sinapov, Matthias Scheutz
International Conference on Autonomous Agents and Multiagent Systems (AAMAS), 2021
pdf

We present a formal framework and implementation in a cognitive agent for novelty handling and demonstrate the efficacy of the proposed methods for detecting and handling a large set of novelties in a crafting task in a simulated environment.

A Framework for Creative Problem Solving Through Action Discovery
Evana Gizzi, Mateo Guaman Castro, Wo Wei Lin, Jivko Sinapov
Workshop on Declarative and Neurosymbolic Representations in Robot Learning and Control, Robotics: Science and Systems (RSS), 2021
pdf

We introduce a unified framework for creative problem solving through action discovery. We describe two methods which enable action discovery at a declarative and neurosymbolic level, namely through action primitive segmentation, and behavior babbling, respectively.

Creative Problem Solving by Robots Using Action Primitive Discovery
Evana Gizzi, Mateo Guaman Castro, Jivko Sinapov
International Conference on Development and Learning (ICDL), 2019
pdf / IEEE

We describe a method for discovering new action primitives through object exploration and action segmentation, which is able to iteratively update the robot's knowledge base on-the-fly until the solution becomes feasible.



Website template from Jon Barron.