Analyzing structural data using deep learning methods

Project ID
FG-573
Project Lead
Project Categories
Computer Science
Abstract

In many scientific domains, structural data is common in many problems. From problems dealing with sequence data such as speech recognition and natural language processing, to problems involving spatial data such as ice-sheet segmentation, to spatial-temporal data such as earth-quake forecasting. There are tremendous values in improving the solutions to these problems using modern technologies. Deep learning has demonstrated great success and huge potential in solving vision, audio, and text related problem. In this project, Dr. Geoffrey Fox and I aim to systematically study the intersection of deep learning, problems involving structured data, and distributed systems. 

Use of FutureSystems
We will use FutureSystems' computing resources to perform experiments with Deep Learning models for analyzing structural data problems in speech recognition, earthquake forecasting, etc.
Scale of Use
We would like to use a GPU node with 8 GPUs (e.g. K80) for 6 months.