Friday, January 24, 2020
PhD candidate Amine Mhedhbi is one of 10 recipients across North America and the only recipient from Canada to receive a 2020 Microsoft Research PhD Fellowship. Established in 2008, these prestigious fellowships have supported more than 150 exceptional doctoral students to date.
Microsoft Research PhD Fellowships are awarded for two years to doctoral candidates in their third year of study. In addition to a $42,000 USD stipend for each of two years, recipients are invited to interview for a salaried internship with leading Microsoft researchers who work on projects related to the recipient’s field of study. As well, recipients are invited to Microsoft’s PhD Summit — a two-day workshop in fall 2020 held at one of Microsoft Research’s labs, where fellows will meet with Microsoft researchers and other top students to share their research.
Supervised by Professor Semih Salihoglu, a member of the Data Systems Group, Amine’s research focuses on understanding and developing new techniques across the graph database stack to improve execution performance of graph analytical workloads. One of his recent projects explores the interplay between worst-case optimal join algorithms and traditional join processing under a hybrid cost-model. In addition, he is working on designing flexible storage for faster graph traversal and using factorized processing to improve graph database query executors.
“Amine leads an ambitious research project to develop a new in-memory graph database management system developed from scratch we call GraphflowDB,” said Professor Salihoglu. “Graph databases are software systems to store, query, and manage graph-structured data — the kind that’s typically thought of as networks. Such data appear in many areas, from social networks to financial networks to data centre traffic networks.
“Amine’s research focuses on two primary questions — how to perform very fast joins to detect complex patterns in graph-structured data, and how to scale an in-memory graph database system. For fast joins, he has worked on integrating fast multiway intersection-based join operators that seamlessly work with traditional join operators as well as operators that can factorize intermediate query results. To scale the system, Amine and two master’s students have been developing a compressed list-based indexing sub-system that allows users to benefit from GraphflowDB’s fast join capabilities for very complex queries, such as those that can detect complex fraud patterns in financial networks.
“In our internal benchmarks, GraphflowDB outperforms existing systems by several orders of magnitude on many workloads, demonstrating the practicality and efficiency of the techniques Amine has been working on. His work has been exceptional and I am very happy to see him get this fellowship.”
“Congratulations to Amine on receiving a prestigious 2020 Microsoft Research PhD Fellowship,” said Mark Giesbrecht, Director of the David R. Cheriton School of Computer Science. “He has also been a recipient of a David R. Cheriton Graduate Scholarship and now with this fellowship from Microsoft he will be able to further expand his exemplary and award-winning research on graph databases.”
The other nine doctoral students selected for the 2020 Microsoft Research PhD Fellowship are from Carnegie Mellon, Georgia Tech, Harvard, MIT, Princeton, University of Illinois at Urbana-Champaign, UCLA and Washington State University.
