My research lies at the intersection of network science, data science, and mathematical biology. I develop mathematical and computational methods for analyzing networks and constructing them from data. A major focus of my research is time-varying networks. I develop methods for representing, embedding, summarizing, and comparing time-varying networks and their nodes. Another focus is the analysis of dynamics on static and time-varying networks, which has led to various methods for multivariate time-series analysis from network-science and related perspectives. These include early warning signals for tipping points in networked systems; inference of energy landscapes from multivariate time-series data; and methods for identifying network communities, core-periphery structure, outlier nodes, and network states, among others. I have worked on a variety of applications involving network and related data analysis, often in collaboration with domain experts. My primary application areas are biology and medicine—for example, gene-expression data; functional magnetic resonance imaging (fMRI), electroencephalography (EEG), and magnetoencephalography (MEG) data from the brain; social networks relevant to medicine; and animal social networks. I also analyze other types of network data, including data from social media, online transaction platforms, air transportation systems, citation networks, and water distribution networks.
How did you end up where you are today? (Your research journey)
I began my academic journey studying mathematical engineering (which is roughly equivalent to applied mathematics) in Japan. My PhD thesis focused on computational neuroscience and related chaotic dynamics. During my postdoctoral years, however, I gradually shifted my research toward network science and mathematical biology. As an early-career faculty member in Japan, around 2008, I also began evolving from a researcher focused primarily on mathematical modeling and simulation into one who increasingly incorporated data science. This transition was deliberately gradual: rather than replacing my background in applied mathematics and mathematical modeling, I wanted to complement it. After a decade of collaborating with experts from various domains (around 2018?), I felt confident that I had been successfully integrating both approaches. As a side note, I have held tenure-track/tenured faculty positions on three continents and spent a substantial amount of time on each: 7.5 years in Japan, 5.5 years in the United Kingdom, and in the United States since 2019. This may be a somewhat unusual academic path.
What makes you excited about your data science and AI research?
When I entered network science around 2002, relevant data certainly existed, but the rapid growth of machine learning, large-scale data science, and modern AI largely came later. It has been exciting to see network-based approaches become increasingly powerful as data have expanded in both quantity and diversity. Many emerging datasets can naturally be represented as networks, and many modern data science methods are closely connected to network concepts and can benefit from network science expertise. I find these growing connections genuinely exciting.
What are 1-3 interesting facts about yourself?
– I jog regularly, as many U-M people do.
– I play the piano.