The Relative Size of Big Data

• Activity Number: 392
• Type: Other
• Date/Time: Tuesday, August 6, 2013 : 2:00 PM to 3:50 PM
• Sponsor: ASA
• Author(s): Bin Yu*+
• Companies: Univ of California at Berkeley
• Keywords: CPU ; computation ; memory ; neuroscienc ; genomics ; remote sensing

Abstract

Big data problems occur when available computing resources (CPU, communication bandwidth, and memory) can not accomodate the computing demand on the data at hand. It is of great interest to develop a framework for Big Data that can encompass a discussion of information extraction at various scales of absolute data size, but at similar scales of relative data size when taking into account available computing resources. Such a framework unifies bottle-neck data problems faced by researchers from industry, national labs, and universities.

We will first review existing works on big data and then attempt at forming such a framework of relative size of big data with concepts of information extraction efficiency relative to computing resource availability. In particular, we will demonstrate this framework through interdisciplinary projects from neuroscience, genomics, and remote sensing.