Mikel Zhobro

I joined prof. Martius and the Autonomous Learning group in November 2023 as a PhD candidate. My research focuses on the intersection of vision, robotics, and machine learning. I am particularly fascinated by the recent advancements in deep learning for image/video generation. My aim is to extend these successes into the 3D realm and tackle complex challenges like 4D generation.

Part of my objective is to develop world models directly from videos in the wild, with transferable knowladge applicable to robotics and and related downstream tasks.


News

Publications

Learning 3D-Gaussian Simulators from RGB Videos

Learning 3D-Gaussian Simulators from RGB Videos

Mikel Zhobro, A. René Geist, Georg Martius
Arxiv, 2025

Train simulators from multi-view videos by leveraging a 3DGS particle set for state representation and maintaining an inductive-bias-free dynamic model using the TEM-PTV3 transformer architecture.

Learning with 3D rotations, a hitchhiker’s guide to SO(3)

Learning with 3D rotations, a hitchhiker’s guide to SO(3)

ICML, 2024

Guidance on selecting the right rotation representations on deep learning models.