April 4, 2013
Air Force awards CSU math grant largest in dept. history
About $3 million to research on developement of statistical tools
The Air Force Office of Scientific Research has awarded the math department at Cleveland State University the largest grant in the history of the department.
The grant of about $3 million will be used to fund research on statistical inference from the topology of complex networks led by Dr. Peter Bubenik, associate professor of math. The grant runs through February 2016.
Bubenik credits two internal awards from Cleveland State for helping him secure this grant. The funding from the graduate college helped his research in the area and present his findings in conferences that attracted attention from other scholars and funding agencies.
The research department of the U.S. Air Force handles large sets of complex data and is interested in looking for ways to understand the complexity in networks in many different application scenarios.
“They have tons and tons of data getting in and they’d like to know what’s going on in the world, what they are looking for is in this data,” Bubenik said.
The hardware for collecting big data has improved dramatically in recent years and the need to understand this data is of compelling and immediate concern.
Mathematics is not the only area that will benefit from understanding these complex networks. Researchers in medicine, health sciences engineering and business all have data coming in that they need to understand explained Bubenik.
Bubenik’s research focuses on developing tools that will allow The Air Force, and other organizations, to understand and draw interpretations from the big data. He uses a spatial approach that allows for the data to be interpreted abstractly—i.e. there are geometric shapes to data that give meaning to what is being collected.
Using his area of interest, applied topology, he wants to take analytical tools that exist in theoretical mathematics and adapt them for applied purposes.
Persistent homology is a tool being looked at by Bubenik in his research that applies to geometric features of data at various scales. Bubenik explained the theory by comparing it to looking at a picture. When you stand far a way from a picture you see certain aspects, if you look at it on a closer scale you see different, finer aspects.
“Same thing happens with data, if you look at it on a large scale you see certain features, if you look at it on a small scale you see other features,” Bubenik said.
“The insight of persistent homology are the features that persist at different scales…if it’s at large and small scales it’s something important.”
It’s not just about being able to take in data and understand it, but to be able to analyze the data and apply it to other things. Combining ideas from topology with statistics is a goal of the research, making it easier to understand the complex networks of data being collected by various institutions and fields.
Currently there is simple software available which allow people to input there information and calculate something that will explain their data. Future hopes of the field are to expand this technology to utilize more complex data for everyday users.
“In basic one dimensional data there is software that people who don’t know anything about statistics can push a button and calculate things that will tell them something about the data,” Bubenik said. “The long term goal is that you can do the same thing with this technology.”
The community of applied topology is one that got its start at big name universities like Duke and Stanford. It later expanded to schools such as Penn State, Chicago and Ohio State.
“It’s kind of nice that CSU is now contributing in this field together with bigger institutions,” Bubenik said.