With the rapid evolution of technology, historical concepts like Taxonomy and Ontology have taken on new meanings and hold almost infinite potential for artificial intelligence and machine learning. Following our recent webinar, CUBE’s Head of RegAI, Elliot Burgess, took time to define taxonomy and ontology and explore how they are being used in regulatory change management.
Taxonomy and ontology are essentially members of the same family – they have similarities and differences but are ultimately intrinsically linked.
A taxonomy is a way to classify hierarchical relationships between things in the same class. As Elliot explains, a taxonomy “is a hierarchical structure of parents, children, grandchildren etc. It delineates a hierarchical relationship going down a tree or up a tree of concepts within a class. Taxonomy is used – particularly in our domain of financial services regulatory change management – to represent lists of things that have a hierarchical relationship.” It is relatively rigid in structure and typically only allows for this up-down movement.
An ontology, on the other hand, enables you to generate and maintain relationships of many types and between different objects. It is more flexible and allows for more complex, deeper and sideways relationships between different projects and different classes. Elliot continues “Where a taxonomy is used primarily to represent hierarchical relationships and only hierarchical relationships between entities, an ontology allows for you to generate many more different types of relationships, where you can define and manage within the W3C semantic framework and Web Ontology Language (OWL) or its variants.”
Linking the two to level-up regulatory change management
Historically, banks have chiefly used taxonomy in their regulatory change management processes. Typically, financial institutions generate, build and maintain a number of taxonomies that represent how their business operates. These are then held in a master-spreadsheets or databases.
This system becomes labour-intensive without the use of ontology. As Elliot points out, when a new piece of regulation is published there will typically be an impact assessment. “A new piece of regulation goes through a process where humans are tagging or analysing or assessing that piece of regulation and mapping it against any number of these taxonomic elements. This is traditionally quite a manual process, done by experts in the field of compliance, and is unfortunately a very repetitive and time consuming, costly exercise.”
Much of this time-consuming, labour intensive work can be streamlined with the introduction of ontology mapped over existing taxonomies.
The introduction of ontology ‘levels-up’ regulatory change management processes. With an ontological framework, a financial institution can integrate multiple taxonomies as, maintaining and defining additional relational information.
Elliot adds that the addition of ontological elements means that “when a regulatory change comes in, it only needs to be mapped to one element – for instance – one policy or one control. And then the ontology will do the rest of the inference work for you…Furthermore, if we map concepts within CUBE’s ontology to content within the customers ontology, effectively what you get is an automated, straight-through mapping of any new regulation directly to the customers’ frameworks and taxonomies.”
The result? Reduced time spent on impact assessment, reduced potential for human error and increased efficiency.