I am a Ph.D. candidate at the Lab for Advanced Software Engineering Research (LASER) in the Department of Computer Science at the University of Massachusetts Amherst.
The goal of my research has been to study and develop techniques for improving the quality and safety of human-intensive processes (HIPs). My dissertation work focuses on an approach for real-time process guidance that aims to reduce the number and impact of human errors in HIPs. My research interests also include requirements engineering, specification and validation of formal process models, process modeling languages and exception handling patterns in processes. I have been studying the above in the context of continuos process improvement and I have been modeling and analyzing medical processes as a case study.
There is evidence that important and critical processes (such as medical processes) performed in our everyday lives contain defects and are prone to human errors that could lead to a variety of negative consequences. Despite considerable work and progress in error reduction, human errors are still a major concern for many HIPs. I am working on developing and evaluating a novel approach for real-time process guidance in an attempt to reduce the number and impact of human errors in HIPs. In this approach, what process performers do is compared to what they should be doing as specified by a process definition (also referred to as a process model), a formal representation of the recommended ways to perform a process. Deviations are detected in real time and possible explanations for these deviations are automatically identified.
To be useful for process guidance, a process definition needs to be sufficiently complete in the sense that it needs to contain all or most of the recommended ways to perform a process. It also needs to be correct with respect to a set of domain goals and requirements. Human-intensive processes, however, are complex and creating a sufficiently complete and correct process definition is not trivial. To address the problem of creating sufficiently complete process definitions, I have investigated different methods for eliciting process information from process performers (or other stakeholders) and for validating resulting process definitions. In an effort to address the problem of creating correct process definitions, I have studied the application of static analysis techniques, such as model checking, to ensure that a process definitions adheres to a set of domain requirements.
To be able to support an automated real-time process guidance and rigorous reasoning and analysis about correctness, a process definition needs to be specified in a notation or a language with precisely defined semantics. I am interested in studying process modeling languages that enable such rigorous analysis.
Many errors in processes occur during exceptional/non-typical scenarios. These errors may occur because people are not well-trained to deal with exceptional situations that often add time pressure, or because technology support can actually become a burden if it was not designed with exceptional situations in mind. I am interested in studying exceptional scenarios in processes and how they can be modeled to facilitate the understanding, analysis, and improvement of processes.