Michael is an Associate Professor of Computer Science and leader of the group. His research has two aims: to make data analysis efficient end-to-end (from data management and processing to machine learning and inference); and to harness computing for social applications (e.g., for data-based policy design or web analysis). Previously, he taught at INSA Lyon for one semester (2017) and spent four years (2013-2017) as a postdoctoral researcher at Aalto University. He completed his doctoral studies at the University of Toronto (2013) and undergraduate studies at the National Technical University of Athens. [University profile]
Arpit completed a Ph.D. in Computer Science during 2019-2023; for a summary, see this news article on the University webpage. His doctoral research developed algorithms for network analysis using network embeddings. In more detail, Arpit uses network embeddings for node classification, network summarization, and graph query processing. His doctoral research appeared at top international venues such as VLDB, WSDM, CIKM, and ECAI. [University profile, doctoral thesis]
Ananth is a doctoral student in Computer Science. His research focuses on developing efficient end-to-end data science systems. Ananth's research is based on the premise that computational efforts can often be shared across different data science tasks. For example, when the data are updated incrementally, it is usually beneficial to update previously trained models incrementally, too, while possibly re-using previous training efforts; or, when multiple users perform similar data science tasks on similar parts of the data, some computations may be performed once and shared across tasks, thus saving computational efforts. Automating such optimizations will lead to data science systems that are not only efficient, but more easily manageable and user-friendly. [University profile]
Sachith is a doctoral student in Computer Science. His research focuses on developing learned index structures. In more detail, Sachith develops algorithms to learn patterns in the available data and harness them for optimized data retrieval and query processing for a given data science task. Specifically, in his current work, Sachith builds indexes for multi-dimensional data that are optimized to adapt to an underlying density model for the data. Such indexes have the potential to be more efficient than traditional, model-agnostic index structures, and thus increase the efficiency of many data science tasks. [University profile]