The development of novel models to integrate and describe complex and heterogeneous data will pose new challenges in terms of efficient, secure, and correct implementation.
This research area will study the algorithms, data structures, and programmatic solutions needed to support effective storage schemes for complex data repositories, focusing on:
-Supporting flexible querying and manipulation of complex data: we will study how to represent, index and query legacy data in different formats, which has not been well-explored but is crucial to handle complex heterogeneous data
-Efficiently managing large data repositories: we will exploit the large and growing gap in the access time of memory hierarchy levels by developing compressed data structures that operate within entropy-bounded space; to support data mining operations we will research indexes that break the entropy barrier by discarding the data and offering only aggregation queries.
-Enforcing and verifying correctness and access privacy: we will build upon the body of techniques developed in the areas of programming languages and program verification in order to develop verification techniques for expressive properties of data-centric systems, e.g., functional correctness, differential privacy, and confidentiality.