Publish at January 31 2024 Updated February 05 2024
Since the invention of printing, the problem of classifying knowledge has only increased. With digitization, the amount of data to be classified, produced by a larger and, above all, better-educated human population, has reached unprecedented heights. The memory of our computers bears witness to this: from k's we've moved on to megs, gigs and now teras. For data centers, we're talking... peta and exa. Soon, we'll be running out of Greek letters to describe the sheer volume of data.
All this accumulated data needs to be classifiable if it is to be related and used. Linear classifications à la Deway, despite their subjectivity, may have done the trick in the days of paper-based knowledge, but they are now well beyond their limits.
Various classification systems have since been developed: some technical, by time or structure; others empirical, such as Google's, initially based on the number of incoming referential links to a document, i.e. on usage recognition; others more flexible, also based on usage reality, but more content-oriented; they consider their semantic links, their meaning links. One such system is faceted classification.
This approach to knowledge classification divides subjects into different facets, rather than organizing them in a linear hierarchy. It enables information to be organized in a way that is particularly well-suited to more complex systems.
In this system, a facet is a particular aspect of a subject, and subjects are classified by combining several facets. New facets can be added, or the relationships between facets adjusted, to suit different types of subject or knowledge domain.
For example, a document about a certain car could be classified by combining manufacturer, model, fuel type and year of manufacture. Faceted classification systems can specify the relationships between facets; some facets can be subordinated to others, enabling more complex structures to be defined.
This system was developed in the 1930s by Belgian researcher Paul Otlet in response to the fact that the same knowledge can be classified in several fields simultaneously. This led to the Répertoire Bibliographique Universel (RBU) and the Classification Décimale Universelle (CDU).
This centralized approach was feasible for bibliographic records in a relatively homogeneous academic world, but became technically unrealistic for managing the astronomical quantity of documents brought about by the IT revolution and internationalization. As a result, the CDU is gradually being replaced by less centralized systems that are easier to use and maintain. But it had not said its last word.
Another faceted classification was also developed in parallel in the 1930s by Shiyali Ramamrita Ranganathan and is known as the "Colon classification". Based on a simple logic (personality (the subject), matter, energy, space and time), it has undergone several iterations and is still in use today.
While in the past, the design of a faceted classification system required great expertise and an overview of what there was to classify, today no human being can have such an overview of all the expanding fields of knowledge. But what's new is that artificial intelligences can. Not only can they do it in terms of capacity, but also, based on the semantic models they have developed, they are capable of keeping themselves up to date and evolving their classification.
In this way, an A.I. can compile and estimate the importance of the relationships that concepts establish between themselves, and from there create a faceted classification that is relevant, useful and leads to the information sought quickly.
Here's an example of a simple faceted classification in mathematics asked of an A.I.
Too simple, let's add a few specific fields: economics, chemistry, fractals and history. We can do it for any field.
To measure the scope of this classification method, and to see the extent to which one discipline can be associated with another, I've added the relationship between mathematics and maternity, two fields where we don't spontaneously see connections.
The result is very interesting. The result is specific, concrete areas where mathematics is applied.
Supported by A.I., the faceted classification approach enables us to explore the links between seemingly distant domains, and leads us to an understanding of the interactions between different fields of knowledge that we wouldn't have thought of at the outset, because A.I. can take into account practically all the knowledge it can access, i.e. more than any human could ever absorb.
Let's try it with "oceanography" and "shipbuilding", two domains on the same subject but having little to do with each other.
Many interesting ideas are emerging, and we could go even further.
Let's try it now with "oceanography" and "fungi", two seemingly completely separate fields.
I suspected there were fungi in the sea, but I didn't think there could be so much interdisciplinarity!
A.I.-assisted faceted classification makes it possible to explore the possibilities of interdisciplinarity in virtually any field, and quickly identify the most promising combinations. In addition to the practical classification of interdisciplinary knowledge, it's a great tool for orientation and rough-cutting.
Illustration: eel000000lee - DepositPhotos
You might also like:
Change your personality with a digital coach
If people can get into physical shape with apps, why shouldn't it be possible to get into psychological shape in the same way? Sometimes we're psychologically overweight too.
How to make a house water self-sufficient
The experience of being autonomous, whether it is autonomous in electricity or autonomous in water, gives an impression of freedom and of no longer being dependent on exogenous factors for its supply. Moreover, even if a 100% autonomy is not reached at the ecological level, it is interesting to preserve the resource that is water.
How to easily detect whether a text has been generated by artificial intelligence
The robust results obtained by the researchers illustrate the unique imprint of machine-generated texts through the alterations produced by the machines themselves.
This software that creates interactive stories
Interactivity in storytelling is not a recent phenomenon. However, with today's technologies, it has never been so accessible. An open-source software is becoming more and more popular as it allows for the easy creation of branching scenarios. This is one more resource for scripting a course as well as offering learners the opportunity to develop their own story.
Your customized A.I. specialist
A.I. then becomes a major element of social influence that you'll have to learn to control. Do you want the truth or what pleases you? Institutions with access to a lot of personal data are all in a position to use and abuse these technologies.
