Cross-Domain MDM Integration Using AI-Driven Data Governance: A Case Study In Financial Technology Architecture

Abstract

Knowledge graphs (KGs) have emerged as a popular and efficient way to integrate multi-domain, multi-model, and multi-level data by gathering facts in the form of triplets and transforming them into machine-understandable representations. These architectures are suitable to integrate highly heterogeneous data sources rendering an appropriate semantics to the domain of a specific vertical industry. [1]Traditionally used in the field of semantic data integration, KGs are currently leveraged in Artificial Intelligence to augment machine learning models and boost enterprise solutions with advanced auto-piloting or analytical capabilities based on inference and reasoning activities. Different verticals have developed proprietary and de facto standard knowledge bases that are public, such as DBpedia, Yago, and Freebase. Growing in acclaim for their synergistic combination of SQL and Machine Learning, KGs have recently gained momentum in FinTech. By providing an insight into things’ connections and relationships in the financial industry, they add a missing piece to FinTech’s state-of-the-art, which consists of Anticipatory Cybersecurity, Business Intelligence, Outlier Detection, Forecasting and Planning, and Data Warehouse.

Data Heterogeneity is a main hurdle in building KGs, especially in the FinTech domain, where data is expected to come from internal and external and proprietary and public sources, each having its ownership and quality issues. Thus, data interoperability should be guaranteed in order to funnel data from these sources into a consolidated enterprise Form of Knowledge Graphs. Data Integration is in its heart a Data Governance Discipline and KGs begin their life as systems integrated feed by Enterprise Data Governance engines. Enterprise Data Governance is not only responsible for Data Integration, it is also in charge for Data Catalogues, Metadata Management, and Data Quality Management, which are the activities in the Data Value Chain that contribute with Semantics, Contexts, and Quality Elixir to the integration of FinTech data into Knowledge Graphs.