Microsoft Future Leaders in Robotics and AI Seminar Series: Faith Johnson

Friday, March 8, 2024
2:30 p.m.
Online Seminar

Towards Socially Aware Visual Navigation with Hierarchical Learning


Faith Johnson
PhD Candidate
Rutgers University

Zoom Link

 
Abstract

Visual navigation follows the intuition that humans can navigate without detailed maps. A common approach is interactive exploration while building a topological graph with images at nodes that can be used for planning. Recent variations learn from passive videos and can navigate using complex social and semantic cues. However, a significant number of training videos are needed, large graphs are utilized, and scenes are not unseen since odometry is utilized. We introduce a new approach to visual navigation using feudal learning, which employs a hierarchical structure consisting of a worker agent, a mid-level manager, and a high-level manager. Key to the feudal learning paradigm, agents at each level see a different aspect of the task and operate at different spatial and temporal scales. Two unique modules are developed in this framework. For the high- level manager, we learn a memory proxy map in a self supervised manner to record prior observations in a learned latent space and avoid the use of graphs and odometry. For the mid-level manager, we develop a waypoint network that outputs intermediate subgoals imitating human waypoint selection during local navigation. This waypoint network is pre-trained using a new, small set of teleoperation videos that we make publicly available, with training environments different from testing environments. The resulting feudal navigation network achieves near SOTA performance, while providing a novel no-RL, no-graph, no-odometry, no-metric map approach to the image goal navigation task.
 
Bio

Faith is a fifth year PhD candidate at Rutgers University focusing on the intersection of socially cognizant robotics and computer vision. She is currently researching how to implement socially aware navigation in robots under the SOCRATES NRT grant with her PI, Professor Kristin Dana. Her other work includes visual navigation agents, pedestrian social behavior understanding, and hierarchical agents for navigation. She graduated magna cum laude with her bachelors in Electrical and Computer Engineering also from Rutgers University in 2019, where she was a member of the inaugural class of the Honors College.
 

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.

Audience: Public 

remind we with google calendar

 

April 2024

SU MO TU WE TH FR SA
31 1 2 3 4 5 6
7 8 9 10 11 12 13
14 15 16 17 18 19 20
21 22 23 24 25 26 27
28 29 30 1 2 3 4
Submit an Event