Hear AllyO’s Cofounder Ankit Somani and Senior Information Engineer Michael Baxter discuss ontology. In this conversation, they will cover:
Ontology is complex and there aren’t many straightforward definitions, especially as ontology relates to the HR world. Here is our simplified definition: Ontology is modeling meaning or relatedness using data.
Meaning is what we do or don’t know. Relatedness is how they connect to each other. Data is the representation.
Think about ontology as taxonomy on steroids.
Ontology is great for sectors like HR where there isn’t a lot of big data, and the vocabulary used is relatively small. You likely will see very common language around hiring. Ontology is great for creating a representation that is able to deal with that lack of information, rather than be oversubscribed with tons and tons of data.
Machine learning has been the pop star for artificial intelligence for the last five years. What differentiates machine learning from ontology is the distinction between pattern matching versus modeling meaning.
Machine learning required strong minds to do the math and clean up computational problems so that something unknown can be labeled or classified by a machine through pattern matching. This takes lots of data and training to achieve a recall result that is human-level or better.
With ontology, we’re doing a knowledge representation that models meaning rather than pattern recognition. So we have the ability to make an inference with meaning when using ontology. This is vital in language where comprehension is strengthened by both the presence of words and the absence of words. For example, “mobile equipment” can mean one thing to an RF engineer and something entirely different if you work in construction.
Cross modalities, or looking at one piece of information and comparing it to another, is one compelling use case for ontology. For an HR use case, think about comparing two cover letters or resumes, and then combining that with additional information collected at different stages of the hiring and employment process. We have the potential to create better, more advanced screening automation.
AllyO filed for a patent for ontology-based systems and methods for HR and recruiting in 2019. To learn more about our technology, demo our solution and speak with one of our experts.
Stay up-to-date with the latest insights and trends from AI recruiting brought to you by AllyO Blog!