Query Paradigms for Non-Traditional Database Applications

Description

Many modern database applications exhibit characteristics that deviate from traditional data management. While traditional data management has focused on “small data” of good quality, modern database applications often have to deal with “big data” that is defective in different ways (incomplete, inconsistent, erroneous). This shift in characteristics raises new challenges that are not adequately addressed by traditional query paradigms. Therefore, new query paradigms are being developed in the database research community. Such developments include consistent query answering (CQA) and approximate query answering (AQA). In CQA, the aim is to provide consistent answers to end-users’ queries even if the underlying data is inconsistent. In AQA, the aim is to provide approximate answers whenever the exact answer cannot be computed in reasonable time (because the computation is intractable or because data is “too big”). These paradigms may also need to be combined. For example, if CQA is intractable, which is known to be the case in many situations, techniques from AQA may be used to approximate consistent query answers. Today, these new query paradigms are immature and their interplay is not well understood. The proposed research aims at contributing to the development of solid theoretical foundations for future database applications.