If we ask a computer to translate the English sentence “the box is in the pen” into other languages, it will most likely interpret the word “pen” as the object we use to write with, this being the more frequently used meaning. But then the sentence will be nonsensical because, as we know, a larger object cannot be inside a smaller one.
Language processing or natural language processing is a much bigger challenge in AI than, for instance, image processing. We humans realize that, for this sentence to make sense, the word “pen” must mean a small area surrounded by a fence. A computer, on the other hand, lacks contextual knowledge and thus the logical reasoning needed to translate the sentence correctly. Another example would be “John is flying to the Big Apple on Tuesday.” You can probably guess what the result would be.
This is where semantic technologies come in. Among the many available methods, semantic techniques aim to improve computers’ understanding in processing natural/conversational languages through knowledge representation. Semantic technology is powered by ontology: it relies on semantic information encoded in ontology to identify nodes (e.g., words) that are semantically related.
At JANZZ.technology, we offer superior semantic technologies including semantic extraction, searching and matching powered by our comprehensive ontology in the domain of occupation data. To illustrate, JANZZ.technology’s semantic solutions can realize the following smart applications:
-Job searching and matching on related concepts
Related concepts are not (necessarily) synonyms but concepts which share similarities, sometimes given in completely different words or even languages. For example, “Neonatology” and “Pediatrics” are related concepts. With the information stored in ontology, semantic technology can identify how closely these two terms/professions are related to each other and, importantly, what kind of training/certifications one of these professionals needs in order to perform the other one’s job. This can be extremely helpful when transforming workforce skills on a large scale such as public employment services.
As another example, “Creative Director” and “Web Designer” are also related concepts but to a much lower degree compared to “Neonatology” and “Pediatrics”. If you are looking for a “Web Designer”, our semantic technologies would also recommend someone with job title “Creative Director” combined with skills in CSS, HTML and UX, or suggest such skills. Of course, “Concepteur Web”, “Nettdesigner”, “مصمم على شبكة الإنترنت” or “网页设计师” will also be matched. Related concepts can also be skills or education. For example, if you are looking for someone experienced with ERP systems, our semantic technologies know that candidates whose CVs list SAP, JD Edwards and MS Dynamics are all good matches because these are all ERP systems.
– Job searching and matching on degrees of skills
Semantic technology is not only able to match job postings and CVs containing the same skills, but it can also compare the degree of skills. For instance, “MS office skills” is a broad term and listed in many CVs. If you are looking for a Spreadsheet pro, you don’t want to be matched with a myriad of CVs listing basic MS office or beginner’s level Excel skills.
Similarly, if you are searching for professional CAD software skills, our semantic technologies would match CVs with CATIA, OpenSCAD or Rhino rather than TinkerCAD or BlocksCAD because the different specificities of CAD software are also stored in our ontology. Moreover, our semantic techniques not only identify levels of skills, but also report any training necessary for candidates to transform skills from one CAD software to another.
– Concept identification through interpretation of the context
Semantic technologies help identify cryptic concepts through context. Job titles can be very challenging for computers to identify. In the sentence “Company X is looking for an RF System Engineer, Building 8, Menlo Park, CA,” our software is able to decode each part of the sentence with the information stored in our ontology, such as industry codes, company names and places of work. In this case, “Building 8” is not an address but instead a mysterious department for hardware development at Facebook, and the “RF System Engineer” refers to “Senior Radio Frequency Engineer”.
– Job matching on overall dimension of occupation data
Some job titles, such as dentist, pilot, carpenter and Android app developer, already contain a lot of information about the specific position. When matching these jobs, it is possible to match almost exclusively on job titles. However, other titles like teacher, consultant, assistant, engineer and coordinator are much less specific. In such cases, one needs to include other criteria such as industry, skills, education, experience, etc., in order to conduct an accurate and meaningful matching. Semantic solutions from JANZZ.technology can perform such tasks with the data linked in ontology.
– Identifying gaps in the information
In contrast to machine learning, which is proficient in pattern recognition and classification, an ontology models meaning. It helps a system to understand CVs and job postings and perform gap analyses, thus creating a more user-friendly experience. For example, when matching candidates and jobs, semantic technologies can recommend skills, education or training which a given candidate lacks and thus help candidates optimize their CV.
Are you a large international corporation, organization or public employment service? Do you want to have the right technology to prepare and accompany your labor force throughout the digital transformation? Do you want to improve user experience during the application process? Do you want to build a more powerful system which makes your products stand out from the HR tech crowd? To integrate the latest semantic extraction, searching and matching technologies powered by JANZZ’s ontology, please write now to firstname.lastname@example.org and let JANZZ.technology assist you.