AI is Taxonomies and Ontologies Coming to Life
Saturday, May 11, 2019 at 10:49AM
Charlie in Becoming an XBRL Master Craftsman

PWC makes the following statement in this article: "artificial intelligence (AI) has the power to change almost everything about the way they do business". You might ask, "How?"

In her presentation, AI and Machine Learning Demystified, Carol Smith makes the following statement on slide 12:

"AI is taxonomies and ontologies coming to life."

I could not agree with her more. But I would say this a bit more precisely. It is not that just any old thing called a "taxonomy" or "ontology" is what you need. You want formal, machine-readable, highly expressive taxonomies and ontologies. Artificial intelligence (AI) works by providing context that the AI can make use of.  Taxonomies and ontologies are about classification. Classification turns unusable "dark data" into knowledge and understanding.  While sometimes it is possible to let machines create classifications using clustering techniques of machine learning; for complex knowledge domains, like financial reporting, humans need to create the classifications. Once you get a solid foundation, then the machine learning processes can leverage the human created information and you get a hybrid approach to expanding taxonomies and ontologies.  Basically, it is taxonomies and ontologies that enable the creation of machine-readable knowledge.  For example:

(Click image for larger view)

And so that is exactly why I created the framework that I did. It is that knowledge that enables Pesseract to do what it does. This is what is meant by the statement "stronger expressiveness and therefore reasoning capabilities" per the ontology spectrum.

Using the thick metadata layer provided by the framework, a business rules processing engine, and mechanisms to present information in human readable form things like this and this and this can be created. Essentially what you can get is similar to the logic and rules-based approach of imlementing artificial intelligence that is similar to the functionality of TurboTax. (Does not work exactly the same, TurboTax is a form; but a financial report is not a form.)

Over the coming years as people see more and more software effectively automating tasks they will come to learn that AI will have a significant impact.  PWC points out that most business leaders have no clue how to implement AI within their organizations. This ignorance on the part of business leaders will lead to wasted money when the snake oil salesmen come around and try and sell you "AI".  It will also result in many business leaders acting either too soon, too late, or going the wrong direction when trying to adapt to this inevitable, and immanent, new paradigm.

"Know how" is a type of knowledge. PWC offers some excellent advice that will help you and your organization maximize their ROI or AI.  Where do you start to turn ignorance into know how? Computer Empathy is the summary that I created from my lab notes.

If you are still having a hard time getting your head around all this, check out this video about how self driving cars work. Or this video.

Article originally appeared on XBRL-based structured digital financial reporting (http://xbrl.squarespace.com/).
See website for complete article licensing information.