2015 Grand Challenge Awardees: ICES Professors Philip Varghese, Rachel Ward, Keshav Pingali, Nichole Rylander, Michael Sacks.
Six faculty received ICES’ 2015 W. A. "Tex" Moncrief Grand Challenge Awards, based on their highly compelling research proposals related to the Grand Challenges in computational engineering and sciences that affect the competitiveness and international standing of the nation.
Aaron Baker, assistant professor of biomedical engineering; Keshav Pingali, professor of computer science; Nichole Rylander, associate professor of mechanical engineering; Michael Sacks, professor of biomedical engineering; Phillip Varghese, professor of aerospace engineering and engineering mechanics; and Rachel Ward, assistant professor of mathematics will receive stipends of up to $75,000 per award per semester to cover salary and other expenses necessary to further their research.
In the traditional paradigm of vascular devices, a medical device company mass produces devices of fixed size and structure that are used in millions of patients. Baker’s research aims to treat occlusive vascular disease by creating patient-specific stenting technology that can be applied to address the geometric and pharmacological needs of an individual patient’s lesion. They will create a set of computational tools to simulate the biomechanical and transport phenomenon during therapy with a drug eluting stent. This toolset will allow the rapid design and manufacturing of custom-designed stents that are engineered to patient-specific arterial structure, hemodynamic flow, and risk of disease.
Pingali’s group develops programming systems that simplify the job of applications programmers, such as computational scientists, who want to exploit parallel programming platforms without having to become expert parallel programmers. Using his Grand Challenge award, Pingali plans to work specifically with machine learning experts to understand their application area, and figure out how to build systems that will be used by their community to exploit parallelism without becoming parallelism experts.
Rylander will develop a computational model to accurately predict the evolving and complex nature of cancer tumor growth, and the response of tumors to candidate therapeutic regimens for treatment optimization. This model will be formulated, calibrated, and validated using a series of physiologically representative in vitro tumor platforms of varying complexity developed by her lab which have been shown to recapitulate the intricacies of the cancer microenvironment.
The primary objective of Sack’s project is to develop novel computational biomechanical models to investigate the biophysical mechanisms that link myocardial infarction to mitral regurgitation. Mitral valve regurgitation is the most common form of heart valve disease in urban Western society with high prevalence across a wide range of age groups. This dysfunction is manifested by the inability of the mitral valve to form a closed seal, which subsequently leads to abnormal blood flow from the left ventricle back into the upper left atrial chamber (i.e. regurgitation), and compromises cardiac performance.
Chemists and engineers have long wanted to calculate the minimum energy shape of molecules because these measurements would aid in computing the rates of a chemical reaction. Varghese’s project seeks to partially answer that question by determining the level of accuracy and location for measurement necessary for describing molecular shapes. Recently he developed an enabling technology that allows him to address these questions and answer them quantitatively for the first time. Previously, researchers could not frame these questions because they could not conceive of running the computations to answer them in a reasonable time. However, his newly developed computational program will run very large numbers of quasi-classical trajectory calculations very efficiently.
Improved medical imaging is one goal for Ward’s work. Here physical constraints impose limitations on the number and type of measurement options, while a strong need remains to achieve accurate reconstruction using these limited number of measurements. How can one combine available, but limited information with additional information given by a class of training images to devise principled strategies for acquiring such images from limited and possibly corrupted measurements? Dr. Ward aims to address these questions using techniques from sparse recovery, statistics, and convex optimization.