Robot Teaching and Learning with Everyday Users in Mind
ECE 595 Seminar Series
September 20, 2024
11:00 AM - 12:00 PM
Location
online via Zoom
Calendar
Download iCal FileRobot Teaching and Learning with Everyday Users in Mind
Presenter: Maram Sakr, University of British Columbia & Monash University
Abstract: Robots have the potential to enhance our quality of life, not by replacing humans, but by collaborating with them to achieve shared goals. Establishing seamless interaction between humans and robots is crucial to this vision. Learning from Demonstration (LfD) allows robots to learn new tasks by observing user demonstrations, opening the door for everyday users to teach robots without needing explicit programming. However, the quality and quantity of these demonstrations directly affect the robot's learning and generalization. Everyday users, while experienced in tasks, often lack the robotics expertise needed to produce high-quality demonstrations, which is a challenge for traditional LfD techniques that rely on expert data. Sakr's research addresses this by understanding the robot’s learning needs to ensure demonstrations meet these needs, training and guiding users to provide informative demonstrations, especially in uncertain areas of the task space, and designing intuitive interfaces that facilitate smooth collaboration between humans and robots. In this talk, Sakr will demonstrate how identifying the robot’s learning needs and guiding users significantly enhance robot learning. Furthermore, I will discuss how these findings point to new directions for rethinking human-robot interfaces, ultimately enabling seamless interactions between robots and everyday users.
Speaker bio: Maram Sakr is a PhD candidate in the joint program between the Mechanical Engineering Department at the University of British Columbia and the Department of Electrical and Computer Engineering at Monash University, advised by Professors Elizabeth Croft, Dana Kulić, and Mike Van der Loos. She was also a visiting scholar at the Intuitive Computing Lab at Johns Hopkins University. Sakr is an experimental roboticist working at the intersection of robotics, machine learning, and human-robot interaction. In particular, she focusses on enabling everyday users to intuitively and efficiently collaborate with and teach robots. Before pursuing her PhD, Sakr earned her master’s degree from Simon Fraser University, Canada, and her bachelor’s degree from Mansoura University, Egypt. Sakr is the recipient of the Alexander Graham Bell Canada Graduate Scholarship from the Natural Sciences and Engineering Research Council of Canada (NSERC) and was the top ranked student in the Computing Science Committee nationwide. She was selected as an HRI Pioneer in 2022 and an RSS Pioneer in 2023.
Faculty Host: Dr. Shuo Han (hanshuo@uic.edu)
Please contact ECE student affairs for login information.
Date posted
Sep 18, 2024
Date updated
Oct 4, 2024