As advances in data collection lead to unprecedented analysis challenges, data science is becoming increasingly important in science, engineering, and business. While there is no single, agreed-upon definition of “data science,” one view puts it at the intersection of computational and inferential thinking.
Regardless of its interpretation, data science is undoubtedly a hot topic within the SIAM community. Themes at recent SIAM annual meetings have included machine learning and statistics; big data, data science, and privacy; integration of models and data; data-enabled modeling and simulation; data assimilation and mining; data science in image reconstruction, processing, and visualization; computational neuroscience; and mathematics of dynamic networks. The SIAM International Conference on Data Mining has continued to attract top papers in the field for over a decade. The SIAM Workshop on Network Science arose organically in 2013 and has attracted more attendees every year. More and more SIAM members, many of them students, are entering the field. The SIAM Activity Group on Data Mining and Analytics has nearly 1,000 members — a fourth of whom represent industry. It is the fourth-largest activity group overall and second-highest in student members.
The new SIAM Journal on Mathematics of Data Science is currently taking submissions.
Data science is also gaining prominence in the field of applied mathematics at large. The National Science Foundation (NSF) report on its 2016 “Theoretical Foundations of Data Science: Algorithmic, Mathematical, and Statistical” workshop
discusses many areas of interest to SIAM, including computational statistics, randomized linear algebra, signal processing, graphs analysis, nonconvex optimization, and multimodal data. The workshop led to the recently-launched NSF program called Transdisciplinary Research in Principles of Data Science (TRIPODS). Various applied math institutions—including the Institute for Mathematics and Its Applications (IMA) at the University of Minnesota, the Institute for Computational and Experimental Research in Mathematics (ICERM) at Brown University, the Statistical and Applied Mathematical Sciences Institute (SAMSI), and the Alan Turing Institute in London—have hosted data science workshops focused on the development of tools and resources to aid research and collaboration.
Recognizing the expanding role of applied mathematics in data science, SIAM has launched the SIAM Journal on Mathematics of Data Science (SIMODS). The National Academies’ 2013 Frontiers in Massive Data Analysis states that, “The research and development necessary for the analysis of massive data goes well beyond the province of a single discipline, and one of the main conclusions of this report is the need for a thoroughgoing interdisciplinarity in approaching problems of massive data.” SIMODS creates a unique opportunity to strengthen the mathematical constituency’s role in the ascent of data science while positively reinforcing the field’s connections to complementary communities in statistics, computer science, network science, and signal processing.
In alignment with SIAM’s traditional strengths, the journal will focus on topics such as numerical algorithms, optimization and control, functional analysis, and theoretical computer science. It will also include additional areas of relevance like machine learning, signal processing and information theory, applied probability, network science, and statistical inference. As with many other SIAM journals, accepted papers may include some combination of algorithm development, scalable computational methods and implementations, and theoretical analysis. Because of SIAM’s strong focus on applications, SIMODS will also feature papers on applications of mathematical methods in data and information processing to science and engineering problems. Interdisciplinary papers are highly encouraged.
Publishing mathematical data science research in one journal, rather than dispersing it across many different journals, will help researchers keep track of the field’s latest developments. SIMODS will provide a natural home for data-focused papers, such as those on mathematical data analysis in the life sciences, relational graph mining in the social sciences, compression of scientific data, data imputation for climate observations, theoretical analysis of machine learning methods, and so on. The journal will thus help SIAM stay relevant to the growing number of current and future members conducting research in this area.
Editor-in-chief Tamara G. Kolda (Sandia National Laboratories) and section editors Alfred Hero (University of Michigan), Michael Jordan (University of California, Berkeley), Robert D. Nowak (University of Wisconsin-Madison), and Joel A. Tropp (California Institute of Technology) have assembled a distinguished team of associate editors with a wealth of experience in applied mathematics, computer science, statistics, signal processing, and network science. They include longtime SIAM members and new colleagues from related fields.
SIMODS is now open for submissions — submit to the journal today!