Research 2009 - 2011
The Minds of Machines
Sheela Ramana, Ph.D.
Professor, Applied Computer Science
Building complex software systems used by banking systems, air traffic controllers, or in aerospace, is a highly collaborative process. Gathering all the requirements for the system from the large number of people involved is a huge challenge. Inherent in any collaboration of such magnitude is technical conflict. Social conflict is another factor: opinions vary, perspectives oppose, interpretations differ, and this can lead back to further technical conflict. Through a combination of computing algorithms and mathematics, and applying models of learning, software engineer Dr. Sheela Ramanna, develops systematic methods for detecting and evaluating these conflicts for resolution.
The first step, collecting the technical requirements for the system, is the most complex, with the greatest margin for error. In fact, designing and writing the software is generally less problematic than collecting all the requirements. Technical specifications may be narrative or mathematical, presenting another challenge: quantifying qualitative information. A framework is created for evaluating the requirements, identifying conflicts and negotiating resolution.
Always deeply interested in machine learning and intelligent systems, Dr. Ramanna is also collaborating with other researchers across the country on a strategic research initiative in bioengineering. Together they are creating a telegaming system (ATA) that provides an automated tracking and assessment for rheumatoid arthritis (RA) of the hand. The goal is to build standardized assessment tool and her role is in applying machine-learning algorithms to reveal the bio-patterns that can provide patient information for analysis.
Using a knowledge-based approach to tracking and assessing the function and impairment of people with RA, they are developing the hardware and software for an electronic game assessment tool. As patients play the game the hand-held module measures various features such as hand movements, strength, stiffness, flexibility, and interaction capabilities with the game. Health practitioners then analyze ATA information to determine progress and treatment.
“For me, this is exciting because I can see how to apply these algorithms in analyzing patterns in people’s functionality. The idea is to provide a basis for objective evaluation for finger/hand function rather than the typical subjective evaluation currently used. I am thrilled we have a graduate program so our students can be working on some very interesting projects.”