The future brings uncertainty.  Predicting what will happen in a given situation—or multiple situations—is often difficult, especially when there is a wide margin for human and natural error.  This is particularly challenging in financial decision-making, such as investment or environmental planning.

To reduce this error, Akhil Langer is developing a parallel computing framework to simplify the decision-making process by incorporating future uncertainty to make decisions that, for example, maximize profits and minimize expenses.  Problems arise when attempting large-scale decision-making on a single computer, because there are too many possible scenarios to process.  This is why Akhil is working on parallel solutions that can exploit tens of thousands of computers to solve large optimization problems.

One application is the problem of aircraft allocation for the U.S. Department of Defense.  The Department of Defense must allocate 1,300 U.S. military aircrafts on different missions—a task it presently assigns manually.  Akhil’s goal is to optimize aircraft allocation using parallel computing on some of the most powerful supercomputers.

As an undergraduate, Akhil previously worked on a project that involved the distribution of health-related information to users through their cell phones.  Users could ask questions, such as how to locate a doctor, or request information on a certain drug via informal text messages.  Akhil's software would then analyze their queries and return appropriate data from a health database to answer their questions.

Akhil received his undergraduate degree in Computer Science from the Indian Institute of Technology Roorkee and holds a Master's degree in Computer Science from the University of Illinois at Urbana-Champaign.

The human genome sequence contains over three billion letters.  Understanding what the vast majority of these letters encode for has long been a mystery.  Emerging technologies are now providing a variety of data on the human genome, but can produce tens of millions of data points—and that’s where computational biologists step in.

Jason Ernst uses computational methodologies to integrate different data sources to analyze the human genome with the ultimate goal of better understanding and treating disease.   He develops and applies computational methods to the data, which can then be used to provide insights into different cell phenotypes and disease-associated DNA variations.

Jason was originally trained as a computer scientist, but became motivated by questions in biology, while a Ph.D student at Carnegie Mellon University studying machine learning.  

The technology and data resulting from the Human Genome Project prompted Jason to recognize that this was a situation in which he could apply machine learning.  He enjoys the computational challenges and opportunity to collaborate with experimental biologists, and hopes genomic work will eventually have a greater effect on personalized medicine and human health. 

Jason was a postdoctoral fellow at MIT during which time he was affiliated with the Broad Institute—a genomic medicine research center in Cambridge, Massachusetts.  He holds his undergraduate degrees in Computer Science and Mathematics from the University of Maryland, College Park, and is currently an Assistant Professor at the University of California, Los Angeles.  In his spare time, he enjoys running, Ultimate Frisbee, and soccer.

The New York Times has hailed Rayid Ghani as Obama’s “secret engine” for re-election.  As Chief Data Scientist, Rayid Ghani invented algorithms to help target voters about the presidential candidates. Ghani and the analytics team broke down the goal of 270 electoral votes into problem sets to answer questions like: who are the swing voters, how do you target each swing voter specifically and how do you mobilize your voters to the voting booth on Election Day?  His team’s work resulted in a list of tens of millions of targeted names and a strategy to optimize their volunteers and funds in the most efficient and effective way. He targeted young voters by encouraging them to sign into the Obama campaign website through their facebook accounts and accessed their social networks to identify persuadable friends. He then encouraged Obama voters to share their Obama pitch with their ten most persuadable facebook friends.

When Associate Professor Jason Hong’s blackjack app on his smartphone asked him for his location, he wondered what his location had to do with a poker game and debated giving the app what it wanted. His curiosity led to an investigation to find out which other apps access personal user information. He took his experiment a step further by presenting his findings to users and gauging their reaction and level of awareness.