Диссертации, представленные на защиту и подготовленные в НИУ ВШЭ

Array (Tensor) DBMSs strive to be the best systems for managing, processing, and even visualizing large multidimensional arrays (tensors). It is a young, fast-evolving, and inherently inter-disciplinary area: many core data types in diverse domains are naturally modeled by arrays (tensors). This dissertation presents fundamental theoretical, systems, and practical contributions to the area of Array (Tensor) DBMSs. Namely, we introduce two novel Research & Development directions: physical world simulations and tunable queries in Array (Tensor) DBMSs. We also propose a new formal Array (Tensor) DBMS data model, novel algorithms, approaches, architectural and implementation aspects for operating on multidimensional arrays (tensors), which can exceed the speed of state-of-the-art approaches and technologies by orders of magnitude. The significance of contributions is demonstrated across a wide variety of practical applications and real-world data. This dissertation is based on the results presented at premier international conferences in computer science: VLDB and SIGMOD.