SBIR-STTR Award

New approaches to handling streaming data are required. By the time one wishes to index stored scientific data, critical decisions about what to store have already been made. We propose evaluating, extending, and implementing a suite of algorithms that use new indexing and sampling techniques to improve analytical performance on large heterogenous streams of scientific data beyond current state of the art.

Keywords:
Algorithms, Streaming Algorithms, Sampling, Indexing, Cloud Computing, Big Data, Distributed Computation, Machine Learning

We consider large, heterogenous data streams comprised of multiple components that may lack underlying models. The problem is to reliably and automatically decompose these high-dimensional objects into a 'background' of simple patterns and a smaller numb

Keywords:
Adaptive Algorithms, Foreground/Background Separation, Convex Optimization, Data Streams, Distributed Computation