Astronomy and physics sky surveys large hadron

Info iconThis preview shows page 1. Sign up to view the full content.

View Full Document Right Arrow Icon
This is the end of the preview. Sign up to access the rest of the document.

Unformatted text preview: , Mendeley 7. Astronomy and Physics: Sky Surveys, Large Hadron Collider (Citations), Netflix, Web Search, Digital Materials, Cargo shipping at CERN and Belle Accelerator in Japan 3. Defense: Sensors, Image surveillance, Situation Assessment 8. Earth, Environmental and Polar Science: Radar Scattering in 4. Healthcare and Life Sciences: Medical records, Graph and Atmosphere, Earthquake, Ocean, Earth Observation, Ice sheet Radar scattering, Earth radar mapping, Climate simulation datasets, Atmospheric turbulence identification, Subsurface Biogeochemistry (microbes to watersheds), AmeriFlux and FLUXNET gas sensors Probabilistic analysis, Pathology, Bioimaging, Genomics, Epidemiology, People Activity models, Biodiversity 5. Deep Learning and Social Media: Driving Car, Geolocate images/cameras, Twitter, Crowd Sourcing, Network Science, NIST benchmark datasets 9. Energy: Smartgrid 5 6 5 Prof. Kai Hwang, USC, Nov. 25, 2013 A Bigdata Case Study : Current Approach and The Future of Deep Learning Large Scale Deep Learning Current Approach: The largest applications so far are to Large models like neural networks over large datasets are increasingly the top performers in benchmark tasks for vision, speech, and natural language processing. image recognition and scientific studies of unsupervised learning with 10 million images and up to 11 billion parameters on a 64 GPU HPC Infiniband supercomputing cluster. The Future: Large datasets of 100 TB will be processed to One needs to train a deep neural network from a large (>>1TB) corpus of data (typically imagery, video, audio, or text). exploit neural networks. Training a self-driving car could take 100 million images at megapixel resolution. Deep learning shares many characteristics with the broader Such training procedures require customization of the neural architecture, learning criteria, and dataset preprocessing. field of machine learning. It takes high computational throughput and high productivity in research exploration. 7 P...
View Full Document

This note was uploaded on 02/04/2014 for the course EE 599 taught by Professor Povinelli during the Spring '08 term at USC.

Ask a homework question - tutors are online