An Introduction to Ontologies
One of the most vital elements of a successful skills development platform is the offer of comprehensive learning pathways. These pathways help identify the skills a learner specifically needs to focus on – a vastly more effective learning process than using a set curriculum for all learners.
Learning pathways are formed by focusing on the skills within skills that someone already knows, needs to know, and will know in the future. As each broader skill branches out into more specific and granular skills, links are created that show the essential connections in knowledge to enable comprehensive and holistic learning. At Workera, we call this an ontology – but where does the term 'ontology' come from and what should it mean to the world of upskilling?
The definition of an ontology
For thousands of years, until a few decades ago, the term ‘ontology’ only referred to a rather esoteric field of philosophical inquiry – the study of what is, or the study of what and how things exist.
More recently, the term “ontology” has been adopted by information scientists who are concerned with classifying and understanding the things that exist in a certain domain. In fact, this use of the term has gained enough currency to earn a place in our dictionaries. According to Oxford Languages, while ‘ontology’ still refers to the aforementioned branch of philosophy, when we speak of “an ontology”, or “ontologies” in the plural, we are referring to “a set of concepts and categories in a subject area or domain that shows their properties and the relations between them."
“For thousands of years, until a few decades
ago, the term ‘ontology’ only referred to a
rather esoteric field of philosophical inquiry – the study of what is, or the study of what and how things exist.”
Thus, we speak of “an ontology” for “a discipline”: anything from Machine Learning to Cat Breeding to Underwater Basket Weaving. Any discipline – also called a ‘domain’ – could have its own ontology.
The taxonomy of an ontology
Looking closer at the above definition, it becomes clear that we not only use ontologies to answer the question “What information is in this domain?” but also to ask “What are the relationships among the information in this domain?”. Obviously, understanding these relationships requires that we break the information down into smaller units, and classify them. In this we can see that an ontology is really a taxonomy, a comprehensive classification, of the information in a domain.
One simple way of imagining an ontology is to think of a tree. A domain is like the trunk of a tree, which then splits into a few subdomains, the largest of the tree’s branches. Next, each subdomain will contain a few topics, the smaller branches, under which there may even be a few sub-topics – the smallest branches of all. For example: in our ontologies here at Workera, we include skills, the equivalent of the leaves on our ontology tree.
To illustrate an ontology applied to real life, let’s use the domain ‘Automobile Mechanics’. This domain might be split into several subdomains, such as Diagnostics, Vehicle Components, and Regular Maintenance. Next come the topics – for example, under Regular Maintenance we are likely to find Oil Replenishment, as well as Air Filter, Windshield Wiper, and Tire Replacement. Finally, under the topic Tire Replacement, we will need to include the following skills: Use a Car Jack, Engage the Parking Brake, and Implement a Torque Wrench.
Our complete ontology might look something like this:
Why ontologies matter
Ontologies provide a comprehensive map of the information in a domain, allowing us to visualize and understand a knowledge base, and its related skill set, in its entirety. It is little wonder, then, that ontologies have proven especially helpful for educators.
For anyone who wishes to develop a comprehensive course of study and final exam for their students, an ontology is an invaluable resource, allowing the instructor to work systematically through each subdomain, topic, and subtopic in a given domain. By working in an orderly manner, the instructor can be sure that students will walk away having learned all the information in the ontology, and developed every skill they need to apply that information in “the real world."
“The last thing a working machine learning
engineer needs to do is work point-by-point through an ontology in machine learning engineering – something they're likely
to already have mastered.”
In the past, working professionals who wished to upskill may have enrolled in just such a comprehensive course in order to encounter a few new concepts and master a few new skills. But that’s the equivalent of a graduate-level student sitting in on a high school or college class.
Frankly, in today’s high-paced work environment, no one has time for that kind of inefficiency and redundancy. The last thing a working machine learning engineer needs to do is work point-by-point through an ontology in machine learning engineering. Chances are, if someone is a machine learning engineer already, this approach would entail many precious hours listening to lectures on already mastered material.
How upskilling solutions use ontologies
Upskilling software allows leaders to identify the most critical portions of an ontology that a learner has yet to master, and then provide a learning pathway to fill in those gaps. In a sense, upskilling solutions enable the leader’s job to sift out all the redundant information a learner might have encountered in a comprehensive course, leaving only the few lectures that the learner would benefit from hearing.
At Workera, for example, we carry out this “sifting” process by having our learners take an assessment at the outset (as opposed to a final exam at the end of a course) which enables our algorithm to make inferences about a learner’s skills based on assessment results. After identifying a learner’s areas for growth using data from the short initial assessment, we can generate a learning plan adapted to the needs of the individual.
We don’t use ontologies to express what every machine learning engineer ought to know, but rather, to determine what this machine learning engineer has yet to learn. And, as it turns out, that’s a big difference.
As a leader, how do you design ontologies to help upskill your team in cutting-edge knowledge?
Learn how Workera can power digital transformation and produce measurable results across your enterprise. Request a demo here!
