Dr. Vinterbo received his PhD in computer science from the Norwegian University of Science and Technology (2000). He is an Associate Professor in the Department of Medicine Division of Biomedical Informatics at the UCSD School of Medicine.
Machine learning and data mining in biomedicine.
This interest can be said to lie in the intersection of computer science, statistics and biomedicine, with a focus on machine learning, algorithms, and formal, knowledge based methods. In particular I am interested in aspects of complexity in set approximations. Smaller and less complex models are likely to be less costly, both in construction and application, are arguably more robust and applicable, and often exhibit performance not significantly worse than their less parsimonious counterparts. However, a model can be too simple and consequently unable to capture all relevant aspects of a problem.
Methods for controlling disclosure in disseminated personal information.
So called "secondary use" of information collected on individuals can be essential, particularly in the health care/biomedical research domain. However, one of the main barriers to effective use of information and data is concerns about the preservation of individuals' privacy. Many of the decisions we make as a society are based on deliberations where we balance cost and benefit, in particular cost to an individual versus the benefit to society. When it comes to dissemination of information about individuals, quantification of benefits is much more readily available than the quantification of privacy risk. The ability to formally bound risk is not only needed to enable sound decision making on a policy level, but also as a foundation on which trust can be built. I am interested in methods that allow access to useful information in data that provide formally verifiable quantitative guarantees with regard to strong and comprehensive definitions of privacy.
More about what I find interesting here.