Enamul Haque

This project explores a new approach to structured data integration by extending traditional relational models into a graph-based framework. We introduce GRM/SQLP, a more expressive and flexible query language that separates schema design from reference resolution.

The work focuses on improving how heterogeneous data sources are combined, especially when dealing with identity and reference inconsistencies. By leveraging graph semantics, this approach can better represent complex relationships across datasets.

Overall, the project aims to make data integration more scalable, intuitive, and robust.

This research investigates how to maintain reliable references within description logic (DL) knowledge bases, especially under data updates. We develop algorithms and frameworks that address identity resolution challenges, ensuring consistency when data evolves over time.

The work combines theoretical foundations with practical mechanisms to manage references more effectively. A key contribution is improving how knowledge systems adapt to changes without losing correctness.

This helps support more dependable reasoning in dynamic data environments.

In this project, I developed a graph pattern matching system using Golang to study the performance benefits of concurrency. The system, GraphMatch, implements subgraph matching using parallel goroutines and communication channels.

The research evaluates how factors like core count, goroutine scaling, and channel size impact performance in combinatorial problems. Results highlight both the strengths and limitations of parallelism, showing where concurrency improves efficiency and where overhead becomes significant.

This work contributes to understanding how concurrent programming can be applied to graph algorithms.

This project focuses on improving data quality in multidimensional ontological systems by identifying and resolving inconsistencies. We propose a framework that uses logical constraints and weighted repairs to determine the most efficient fixes to problematic data.

By modeling data quality as a context-dependent problem, we leverage Answer Set Programming to compute optimal repairs. The approach considers both data and schema-level inconsistencies, offering a principled way to clean complex datasets.

Ultimately, this work aims to make analytical systems more reliable by ensuring their underlying data is consistent.