Event
Booz Allen Hamilton Colloquium - Amazon Consumer Robotics and AWS Physical Stores Technology
Friday, November 17, 2023
3:30 p.m.-4:30 p.m.
Jeong H. Kim Engineering Building, Room 1110
Darcy Long
301 405 3114
dlong123@umd.edu
Speaker: Darnell Moore, Amazon Consumer Robotics
Talk Title: Meet Astro: Amazon Consumer Robotics' First Robot for Homes and Small-to-Medium Businesses
Abstract: Astro is Amazon’s first robot designed for homes and small-to-medium businesses. In this talk, we'll introduce Astro by examining the mobility and perception systems that enable autonomous navigation in unstructured environments and support understanding of the surroundings. We also highlighting challenges faced during product development and testing, such as design for manufacturability. We'll also feature some of the capabilities we've added to make Astro more user-friendly.
Bio: Darnell Moore, Ph.D. leads academic partnerships for Amazon Consumer Robotics, where he builds collaborations with universities to accelerate the state-of-the-art in strategic domains like AI, robotics, machine learning, and computer vision. Before joining Amazon in 2021, Darnell was a Distinguished Member of Technical Staff and Manager at Texas Instruments, where he lead researchers working at the intersection of silicon architecture and computer vision. Over his career, he has developed hardware accelerators used in millions of devices, been awarded nine patents, published technical articles, and chaired technical conferences and workshops. A Lifetime member of the National Society of Black Engineers (NSBE), he was recognized as the organization’s 2017 Distinguished Member of the Year. Dr. Moore completed a Master’s and Doctorate from Georgia Tech and a Bachelor’s from Northwestern University, all in electrical engineering. A native of Chattanooga, Tennessee, he is a member of Alpha Phi Alpha Fraternity, Inc.
Speaker: Gregory D. Hager, AWS Physical Stores Technology
Title: Building AI-Powered Computer Vision Systems that Customers Can Trust
Abstract: Amazon is revolutionizing the shopping experience by developing ground-breaking products such as Amazon Just Walk Out technology and the Amazon Dash Cart. However, creating these experiences requires us to confront the long-tail challenges of the real world while maintaining the availability and accuracy that our customers have come to expect from existing physical retail systems. Despite the rapid and continued advances in the capabilities of machine learning within the research community, there is still a substantial gap that we have to bridge in order achieve the reliability and accuracy to necessary to earn our customers’ trust. In this talk, I will discuss the challenges and research opportunities for machine-learning powered computer vision systems that operate in the real world, and illustrate some of the ways that the challenges of accuracy and reliability can be surmounted for systems supporting physical retail.
Bio: Gregory D. Hager is a Senior Principal, Applied Science for AWS Physical Stores Technology. His team supports both Amazon Just Walk Out technology and the Amazon Dash Cart. He is also the Mandell Bellmore Professor of Computer Science at Johns Hopkins University (on leave) and the Founding Director of the Malone Center for Engineering in Healthcare. Greg’s research interests include computer vision, vision-based and collaborative robotics, time-series analysis of image data, and applications of image analysis and robotics in medicine and in manufacturing. He has over 500 peer-reviewed publications and 30 patents on these topics. He has served on the editorial boards of IEEE TRO, IEEE PAMI, and IJCV and ACM Transactions on Computing for Healthcare. He is a fellow of the ACM and IEEE for his contributions to Vision-Based Robotics and a Fellow of AAAS, the MICCAI Society and of AIMBE for his contributions to imaging and his work on the analysis of surgical technical skill.
