报告题目：Boolean Matrix Decompositions and Its Applications
报告人：陆海兵 教授 （美国Santa Clara University）
报告摘要：Many real-world data sets can be represented in the form of Boolean (0/1) matrices, e.g. transactional data (purchase or not), product reviews (like or dislike), documents (containing a term or not). Boolean matrix decomposition (BMD) recently has stood out as an effective model for analyzing/mining Boolean data sets. BMD is to decompose a Boolean (0/1) matrix into the Boolean product (one plus one equals one) of two other Boolean matrices. The key advantage of BMD over other matrix decomposition methods is the direct interpretability of its decomposition solutions, e.g. one decomposed matrix can be viewed as extracted patterns and the other composes the source data with the extract patterns. This study intends to build a general framework of BMD and explores its practical applications. Many variants of BMD, such as minimal rank BMD, minimal edge BMD, rank-one BMD, extended BMD with negative ones, are proposed and investigated. Due to the discrete nature of BMD, most of the studied problems are NP-hard. So we design and implement several algorithms, including algorithms with guaranteed approximation bounds, fast and dirty heuristics. We also formulate BMD variants as standard optimization problems that allow data practitioners to directly apply existing optimization solvers. Our research results can be used in many domains, including shopping pattern discovery, product-rating analysis, text mining, and role-based access control in information security.
报告人简历：Dr. Haibing Lu received his Ph.D. in Management Information Systems in 2011 from Rutgers University and earned his B.S. and M.S. degrees both in mathematics from Xi’an Jiaotong University, China, in 1998 and 2002 respectively. Dr. Lu joined the Department of Information Systems and Analytics (formerly Operations Management and Information Systems) in the Leavey School of Business at Santa Clara University in fall 2011. He was promoted to tenured associate professor in 2017, and department chair at the same time. He is a frequent recipient of the school’s Extraordinary Research Award and Extraordinary Teaching Award.
Dr. Lu’s research is at the confluence of data analytics, information privacy, and optimization. He has published over 40 well-cited research articles at leading journals, such as INFORMS Journal on Computing (JOC), IEEE Transaction on Dependable and Security Computing (TDSC), Journal of Computer Security (JCS), OMEGA and Expert Systems with Applications, and premier computer science conference proceedings, including IEEE Symposium on Security and Privacy (S&P), ACM SIGKDD International Conference on Knowledge Discovery and Data Mining (KDD), SIAM International Conference on Data Mining (SDM), IEEE International Conference on Data Mining (ICDM), and IEEE International Conference on Data Engineering (ICDE).