Microsoft Future Leaders in Robotics and AI Seminar Series: Jana Pavlasek
Friday, February 16, 2024
Think graphical, act local: Distributed inference for multi-robot coordination
University of Michigan
To achieve the long-promised general robot assistants capable of operating in our uncertain and unstructured world, we need efficient, reliable, and cooperative AI systems. My research explores advancements in distributed methods for efficient, adaptive, and robust inference for robotic perception and planning problems. Distributed inference leverages graphical models to break up complex problems into multiple smaller subproblems which can be solved in parallel. In this talk, I will present my work on distributed inference for challenging robotic problems including robot perception and multi-robot coordination. I will demonstrate how nonparametric representations of uncertainty help tackle high-dimensional problems with multiple diverse solutions efficiently. First, I will present my contributions to nonparametric belief propagation for the problem of articulated object localization in cluttered scenes towards robot manipulation of hand-tools. Next, I will show how these ideas are extended to distributed planning over multi-robot teams through a new algorithm called Stein Variational Belief Propagation.
Jana Pavlasek is a PhD candidate in the Robotics Department at the University of Michigan advised by Professor Chad Jenkins. Her research interests include robotic perception and planning under uncertainty in challenging, real-world environments. Particularly, her work focuses on scalable distributed inference algorithms, including perception in clutter and multi-robot coordination. Jana is the developer and instructor for a core course on introductory robotics for the Robotics undergraduate degree at the University of Michigan, which has also been taught across several other universities as part of a distributed teaching initiative (hellorob.org). Prior to her PhD, Jana received a Bachelor’s degree in Honours Electrical Engineering at McGill University. She is the recipient of the NSERC Postgraduate Scholarship and has interned at NVIDIA Research. https://janapavlasek.com/
About the Seminar Series
The Future Leaders in Robotics and AI: Celebrating Diversity and Innovation Seminar Series is part of the University of Maryland and Microsoft Robotics and Diversity Initiative. This is a nationwide online seminar series for PhD students, postdoctoral researchers, or early-career professionals, especially underrepresented minorities and women. The seminar series highlights the latest research and innovation in the field of robotics and AI. The series is intended to provide exposure and mentorship opportunities to the speakers, build a network of innovators across the country, and support the speakers’ career planning.
The seminars are held once per month during the academic year. There are two speakers per seminar. Each speaker gives a 20-minute research presentation followed by a Q&A segment. Immediately after the second seminar, the speakers participate in a discussion with faculty.