Modern data analysis increasingly relies on graph-based formulations to uncover structure, detect communities, and guide downstream learning tasks. This talk presents recent advances developed at LAU's Combinatorics in AI Laboratory (CAIL) that demonstrate how classical graph modification problems, particularly cluster editing and bicluster editing, can serve as practical and effective tools for clustering and biclustering high-dimensional and noisy data.
A central theme is the introduction of vertex splitting, a graph-theoretic operation that enables overlapping and flexible cluster assignments while maintaining algorithmic tractability. This framework generalizes traditional editing models and leads to new exact, parameterized, and heuristic algorithms with strong empirical performance.

The talk highlights experimental results showing how these methods scale to real datasets and how they compare to established machine-learning clustering approaches.
The presentation concludes with a discussion of scalability, including the use of parallel graph modification techniques and the integration of algorithmic and data-driven strategies. The talk also outlines open challenges and future opportunities in using graph algorithms to support large-scale clustering, biclustering, and structure discovery in complex datasets.

Dr. Faisal N. Abu-Khzam is a Professor of Computer Science and Assistant Dean for Research Collaboration and Partnerships at the Lebanese American University (LAU). He earned his Ph.D. in Computer Science from the University of Tennessee in Knoxville, and both his M.S. in Mathematics and B.S. in Computer Science from the American University of Beirut. His research interests span graph algorithms, combinatorics, artificial intelligence, and high-performance computing.

Abu-Khzam's research contributions include the development of key methods for parameterized and exact algorithms, record-breaking kernelization bounds, several of the fastest known exact and enumeration algorithms for classical graph problems, and pioneering work on parallel exact algorithms for large-scale combinatorial problems. His recent work includes novel methods for correlation clustering, feature selection, and graph embedding in machine learning. Through his research, he has advanced theoretical foundations while fostering practical applications in areas such as biology and environmental sciences.