Research in Computational Science

As above, computational science is the application of computing and mathematics to interesting scientific problems, where science is defined as "any endeavor of study".  We call this "computational X", where X is anything you might want to study!  For example, you can have computational chemistry, computational biology, and computational physics, but you can also have computational music theory, computational art history, computational linguistics, and computational sport science -- hopefully you get the idea!  Some of the best projects have been the most unusual, for example, the student who studied Egyptian hieroglyphics and the student who investigated the use of facial recognition on race horses to see if they would be competitive on the horse track!  

To do computational science, you use models that others have built, modify models that others have built, and/or build your own model.  In this class, we teach a half-dozen or so computational tools, including things like Mathematica, R, STELLA, Pymol, Paraview, and others.  Some of them require programming, some do not.  We never use computing tools that only have application in education -- all of the tools taught are research-grade, so you will learn how to use tools that can make you a "marketable" commodity to university researchers.  Many of my students get research jobs at the university because they know these tools.  

In this class, you will never work in what you might think of as a traditional lab.  Our experiments are theoretical/computational, done on a computer.  No test tubes, no dead cats—your lab is on your laptop!  Some students need more computing power than is available on a laptop, and those students use the supercomputer at the Pittsburgh Supercomputing Center, where we have an allocation of compute resources.  

In the spring semester, it's all about developing a research question.  To do that, we spend a great deal of time reading the literature, presenting that reading in Journal Club (JClub) presentations, and working to develop a proposal for evaluation.   The proposal follows the style of proposals for the National Science Foundation, so students learn early how science gets funded at the university level. 

Most of the research work is done in the summer, typically through the SRIP program.  Some RCompSci students have external mentors, and some do not.  

Students return in the fall to finish data collection and analysis and then write their final paper and poster.  Many will submit to competitions such as Regeneron STS and Regeneron ISEF.  Students also typically work collaboratively on a group project, such as developing teaching materials for high school students in the computational sciences.  This year, students are developing a “Student Researcher Guide to Computational Tools”, a document that will provide useful information to any research student or instructor who wishes to include computation as a part of their research.

Students in this program must, like all of the RSci programs, be self-motivated, intellectually curious, and good time managers.  Specifically for this program, successful students must enjoy working with computers, but that does not mean writing code!  Some students write code, while others use existing software such as Schrodinger, StarDrop, or Gaussian to do their work.  This is NOT a computer science research program.  Some students need the high-performance computing resources available at the Pittsburgh Supercomputing Center, but most students are able to do all of their research using their laptops!  That is certainly one advantage computational science students have over their experimental peers -- you always have your "lab" wherever you and your computer are together! 

• Computational Fluid Dynamics for Cardiovascular Modeling

• Chinese Natural Language Processing (NLP)

• Computational Drug Optimization of Opioids

• Atmospheric-Marine Systems-Dynamics and Puget Sound Shellfish

• Using Deep Learning for Observational Cosmology

• Computational Neuroscience for Alzheimer’s Disease

• Reinforcement Learning

• Computational Disease Prediction for Predicting Autism

• Facial Recognition of Race Horses

• Computational Analysis of Egyptian Hieroglyphics

• Computational Environmental Nanomaterials

• Medical Image Analysis for Cancer Identification

• Computational Image Processing

• Computational Acoustic Source Identification

• Computational Number Theory

• Cryptography and Data Structures