Iraklis Psaroudakis from Oracle Labs in Zurich will give a talk in the elite program’s special lecture series. The title of the talk is “Introduction to Graph Processing and Analytics (with PGX)” and it will take place virtually as an online meeting at 4:00PM on July 15th 2021.
Graph analytics are one of the top data analytics trends. We begin by explaining why organizations are so keen on modeling data as a graph to express entities and their relationships as first-class citizens. Graphs allow to more easily gain complex insights from data (e.g., retail, healthcare or financial data) through a mix of expressive and powerful graph processing approaches: graph algorithms, querying, and machine learning. We drill down on each one of these three approaches, the differences between graph analytics/algorithms (such as Pagerank ) and graph queries (such as
(:person)-[:friend]→(:person)), how they can be used, and the processing challenges they pose. We describe how Oracle Lab’s  in-memory graph processing framework, Parallel Graph Analytix (PGX) , tackles these processing challenges to develop a high-performance processing solution for large-scale graphs. Next, we continue to show how graphs can be created, modified, queried and visualized using Oracle Labs Data Studio’s interactive visual notebooks. Finally, we focus on a prominent real-world use case: we describe how Oracle’s Financial Crime and Compliance Studio (FCC Studio ) uses graph analytics & visualization to help a bank analyze its data and fight financial crime. We describe how this goal requires solving several technical & research challenges, and how we approach them. One of the major challenges is combining graph, entity resolution, machine learning, and big data techniques to correlate a bank’s internal customer data with external data (e.g., watchlists) into a financial graph, and help investigators detect patterns of criminal activity.
Iraklis Psaroudakis is a Principal Member of Technical Staff at Oracle Labs (Switzerland). His research interests include analytical & graph workloads, parallel programming, OS/runtime-system interaction, machine learning, and financial crime & compliance. Prior to Oracle, he completed his Ph.D. at the Data-Intensive Application and Systems (DIAS) Laboratory of EPFL, Switzerland, focusing on scaling up highly concurrent analytical database workloads on multi-socket multi-core servers through (a) sharing data and work across concurrent queries, and (b) adaptive NUMA-aware data placement and task scheduling. During his Ph.D., he cooperated with the SAP HANA database team. Before starting his Ph.D., he completed his studies in Electrical & Computer Engineering at the National Technical University of Athens (NTUA), Greece.