Why interoperability is vital to sharing data and linking systems – part 1

Raymonde Weyzen

Semantic Interoperability - a much-heard term in the field of data sharing these days but a less easy-to-understand concept. Many systems and advanced technologies, like artificial intelligence (AI), rely heavily on semantic interoperability to create a common vocabulary for accurate and reliable communication among machines 1 . But what is semantic interoperability and how does is affect data sharing between systems?

Let me explain with an analogy. Imagine for a moment that you are walking into a library where all the books have blank covers. How will you be able to find what you are looking for or how can you simply explore authors, topics, genres, or time-periods. Without any information on the cover, that is, without any identifiers, it becomes impossible to pick a book without reading it.

To find out which books are relevant for you, they need to have a number of characteristics. More importantly, the way different people define these characteristics (the meaning) has to match. I might call Homer’s Odyssey “a classic” but someone else might label it “historical”. Agreeing on the labels and the method for labelling is what semantic interoperability in essence represents. And it works the same way with data and code.

A computer, or any type of system that deals with data, needs a method to gauge what a piece of code, or any entry in a dataset, represents. With the growing amount of data that advanced technologies use, and the increasing diversity of this data, agreeing on semantics becomes even more important. However, the way meaning is represented in code is difficult for computers to grasp. Computers don’t understand what the texts they read or write actually mean. Yet with metadata standards, vocabularies, and metalanguage, clear advances have been made. 2 The European Commission’s Semantic Interoperability Community (SEMIC) is a clear example here, having defined core vocabularies, data catalogues, and metadata schema for public administrators. Another example is the Information Economy Meta Language (IEML), which has a compact dictionary of less than 5000 words and organises words by subject-oriented paradigms with a completely regular grammar. 3 The meaning of complex concepts and relations is expressed in this grammar by combining simpler concepts, basically like buildings blocks.

A semantic code like this can support many applications in various domains. In the context of health care, for example, semantic code can enrich data with context and meaning, and improve the understanding of evidence and the interpretation of disease and well-being, not just for patients but also for healthcare providers and researchers. It also offers an opportunity to combine this data with socio-economic information to provide a more multi-faceted picture. Taking this even further, you can think of occupational classifications and international labour market statistics that automatically agree on semantics or government statistics, national libraries, major museums, and digital humanities research that are entirely interoperable. 4 With uniform and specific labels and definitions, AI could become not just more efficient, but also more transparent.

Why interoperability is vital to sharing data and linking systems – part 1
2020, Uriel via Unsplash

Für Fragen und Kommentare besuchen Sie bitte unser Forum auf Futurium.