Senior Editor: Balaji Padmanabhan

University of South Florida

Term: January 1, 2021 – December 31, 2023


Research Background and Interests

My background is computational. My doctoral work in the late 90s was in algorithms for pattern discovery from data. Since then I have continued to work in the data mining, analytics, machine learning, data science and AI areas (as these names kept changing). I have strong interest in new algorithms and innovative new applications in business, health and social contexts. Often the algorithms that interest me are data-driven ones, and they are therefore usually motivated from research in the machine learning area. Other times these algorithms are computational artifacts that solve some important problem. I also enjoy data-driven design, where these algorithmic and computational artifacts come together in a practical system that adds value. In IS building such systems is critical and the design science methodology has a lot to offer here.

I appreciate important algorithmic and computational contributions while at the same time try to ask questions that matter in practice. My desire for relevance has taken me to applications in healthcare, news recommender systems, platform optimization and social and political applications as well. Many years ago I recall a researcher from Bell Labs talking about a visit from a superstar scientist who asked all the researchers in the lab three questions: (a) What are the most important problems in your field? (b) Are you working on those? © Why not? These three questions have stayed with me and I try and encourage people around me to try and reflect in this manner as well.

Methodology

Computational, algorithmic and machine learning methodologies (see differentiation below), along with the broader design science framework, are my primary methodological interests. Machine learning is specifically about learning from data. Algorithmic is more general, in the sense that there are many important algorithms that aren’t data-driven. Agent-based models for one, and many network algorithms as well are examples. Computational (thinking) to me is even more general, and usually results in the design of intelligent systems (including AI today) to solve problems and embraces design science ideas. I use computational here to reflect a systemic view of solving problems, not to reflect the theory of computation (which is closer to the algorithmic view).

Preferences

I prefer to handle papers that are primarily computational and algorithmic as noted above. While often these involve data, there are other theory-driven methods including econometric techniques that are often used with large data to answer causal questions. I am less interested in handling papers that use those techniques primarily because there are others in the editorial board with that expertise. I enjoy shorter papers, where the authors have invested substantial time in ensuring that every sentence is there for a reason, and where there is little to no repetition of ideas. I also enjoy papers that only present the core ideas and results needed to make the point. Oftentimes our own work makes us do extensive analyses, but the onus is on us, as authors, to prioritize and write in a manner that reflects value not effort. I also believe that the onus is on us, as authors, to write in a manner such that our work is understandable. Papers that are overly complex, where reviewers struggle to evaluate correctness and contributions, usually do not do well in the review process.


Email: bp@usf.edu