Live-Stream


Uniting the Computer Vision and Remote Sensing Communities with Open-Source Competitions

05 May  10:00 am EDT

Abstract: The computer vision community has made great strides in the last decade, due in large part to the growth of deep learning. The application of advanced computer vision techniques to the remote sensing domain has lagged considerably, however, due to a number of a factors such a lack of labeled data, disparate object sizes, and the sheer scale of remote sensing datasets. We describe efforts to accelerate the application of advanced computer vision techniques to the remote sensing data via the SpaceNet initiative. SpaceNet is inspired by the highly successful ImageNet competition series that helped spur many of the advances in computer vision applied to traditional imagery. Accordingly, the SpaceNet partners have released a large, high quality dataset containing high-resolution satellite imagery and hand-labeled objects of interest. On top of this dataset, the SpaceNet partners run a series of open competitions that showcase the use cases of computer vision and data science techniques with satellite imagery. We discuss some of the lessons learned from early competitions that focused on foundational mapping techniques such as building footprint extraction and road network detection. Further, we discuss existing capability gaps and plans for future competitions that will address such issues as seasonal variability, robust change detection, and fusion with non-imagery datasets.

Speaker
Adam Van Etten is the Technical Director of CosmiQ Works, an In-Q-Tel Lab. In this role, he applies machine learning and computer vision techniques to satellite imaging data, focused on problems of interest to the US Government. Adam has focused on helping run the SpaceNet initiative, and on researching rapid computer vision techniques that readily scale to the enormous sizes of satellite imagery corpora. Prior to In-Q-Tel, Adam was a Data Scientist for Data Tactics working at DARPA headquarters developing tools and scalable algorithms for big data analysis on a variety of projects. Adam received his Ph.D. in physics from Stanford University and bachelors in physics and astronomy from the University of Washington.

Abstract:

GPU/accelerator architectures have greatly improved the training and inferencing speed for neural-network-based machine learning models. As major industry players race to develop ambitious applications such as self-driving vehicles, unstructured data analytics, human-level interactive systems, and human intelligence augmentation, major challenges remain in computational methods as well as hardware/software infrastructures required for these applications to be effective, robust, responsive, accountable and cost-effective. These applications impose much higher levels of data storage capacity, access latency, energy efficiency, and throughput. In this talk, I will present a vision for building a new generation of computing components and systems for these applications.

Speaker’s Bio:

Wen-mei W. Hwu is a Professor and holds the Sanders-AMD Endowed Chair in the Department of Electrical and Computer Engineering, University of Illinois at Urbana-Champaign. He is the director of the IMPACT research group (www.crhc.uiuc.edu/Impact). He co-directs the IBM-Illinois Center for Cognitive Computing Systems Research (C3SR) and serves as one of the principal investigators of the NSF Blue Waters Petascale supercomputer. For his contributions, he received the ACM SigArch Maurice Wilkes Award, the ACM Grace Murray Hopper Award, the IEEE Computer Society Charles Babbage Award, the ISCA Influential Paper Award, the IEEE Computer Society B. R. Rau Award and the Distinguished Alumni Award in Computer Science of the University of California, Berkeley. He is a fellow of IEEE and ACM. Dr. Hwu received his Ph.D. degree in Computer Science from the University of California, Berkeley.

Description

The Low Power Image Recognition Challenge (LPIRC) 2019 is a one-day workshop that will extend the successes of LPIRC from the past four years, identifying the best computer vision solutions that can simultaneously achieve high accuracy and energy efficiency. Since the first competition, held in 2015, the winners’ solutions have improved 24x in the ratio of accuracy divided by energy.

The live-stream of LPIRC will feature presentations from researchers and last year's winner (all times PDT):

09:30-09:40 - Welcome by Organizers and Summary of Online Challenge

09:40-10:30 - 2018 competition winners will give a talk on their winning solutions — Amazon's Tao Sheng, and Expasoft's Alexander Goncharenko and Sergey Alyamkin

10:30-11:10 - Invited Talk: Rethinking the Computations in Computer Vision (and the Hardware that Computes Them) - UC Berkeley's Kurt Keutzer

11:10-13:40 - Live-stream will shut down temporarily

13:40-14:00 - Invited Talk: Visual Wake Words Challenge - Google's Aakanksha Chowdhery and Pete Warden

Featured Speaker 1

Title, Company

Gingerbread oat cake powder carrot cake. Lemon drops pie icing. Cotton candy jelly halvah sesame snaps jelly powder danish. Pie sweet roll cake lollipop cake. Icing pie dragée gummies jelly soufflé carrot cake chupa chups chocolate. Cake pie jujubes chocolate cake marshmallow topping tart tiramisu.

Additional Supporting Details

  • Additional Speaker

    Title, Company

    Brownie sesame snaps candy canes. Wafer muffin powder chocolate bear claw bonbon pastry. Topping caramels carrot cake marshmallow soufflé icing.

  • Additional Speaker

    Title, Company

    Brownie sesame snaps candy canes. Wafer muffin powder chocolate bear claw bonbon pastry. Topping caramels carrot cake marshmallow soufflé icing.

  • Additional Speaker

    Title, Company

    Brownie sesame snaps candy canes. Wafer muffin powder chocolate bear claw bonbon pastry. Topping caramels carrot cake marshmallow soufflé icing.

Oat cake toffee chocolate sweet

Muffin croissant fruitcake candy cake.

Gingerbread oat cake powder carrot cake. Lemon drops pie icing. Cotton candy jelly halvah sesame snaps jelly powder danish. Pie sweet roll cake lollipop cake. Icing pie dragée gummies jelly soufflé carrot cake chupa chups chocolate. Cake pie jujubes chocolate cake marshmallow topping tart tiramisu. Tootsie roll cupcake halvah sweet muffin. Bear claw jelly lemon drops dessert dessert.

  • Brownie sesame snaps candy canes. Wafer muffin powder chocolate bear claw bonbon pastry. Topping caramels carrot cake marshmallow soufflé icing.

    > Read More

  • Brownie sesame snaps candy canes. Wafer muffin powder chocolate bear claw bonbon pastry. Topping caramels carrot cake marshmallow soufflé icing.

    > Read More

  • Brownie sesame snaps candy canes. Wafer muffin powder chocolate bear claw bonbon pastry. Topping caramels carrot cake marshmallow soufflé icing.

    > Read More

CTA

CTA