BLOG:  Digital Financial Reporting

This is a blog for information relating to digital financial reporting.  This blog is basically my "lab notebook" for experimenting and learning about XBRL-based digital financial reporting.  This is my brain storming platform.  This is where I think out loud (i.e. publicly) about digital financial reporting. This information is for innovators and early adopters who are ushering in a new era of accounting, reporting, auditing, and analysis in a digital environment.

Much of the information contained in this blog is synthasized, summarized, condensed, better organized and articulated in my book XBRL for Dummies and in the chapters of Intelligent XBRL-based Digital Financial Reporting. If you have any questions, feel free to contact me.

Updated US GAAP Financial Report Ontology Prototype

I have made a bunch of updates to the US GAAP Financial Report Ontology.  Have a look at Topics, Disclosures, Templates, Examples, References, Networks, Classes, Reporting Styles, Fundamental Accounting Concepts, and Consistency.

I am not quite sure what to do with Properties yet.  Still thinking about that.  Aiso, I do need to show a list of Elements; but that is straight forward other than the fact that one needs to deal with the fact that there are 14,000 or so elements in total.

More to come so stay tuned!

Not quite sure why all this is important?  You might want to check out Artificial Intelligence and Knowledge Engineering Basics in a Nutshell.

Posted on Monday, July 22, 2019 at 03:33PM by Registered CommenterCharlie in | CommentsPost a Comment | EmailEmail | PrintPrint

Artificial Intelligence and Knowledge Engineering Basics in a Nutshell

Artificial Intelligence and Knowledge Engineering Basics in a Nutshell summarizes into one place all the things what I wish I would have known 20 years ago when I started working with XBRL.

If you want to become a master craftsmen, you need this understanding.

I have not been able to find a resource that explains artificial intelligence adequately.  I have not been able to find a resource that explains knowledge engineering basics in a manner that is approachable by a business professional.  I have not been able to find a resource that explains the connection between artificial intelligence and knowledge engineering.  This 45 page resource provides all three of those things.

With 50% failure rate of artificial intelligence related projects implemented within the enterprise; you might want to consider investing in understanding this new, important capability.

Posted on Monday, July 22, 2019 at 08:30AM by Registered CommenterCharlie in | CommentsPost a Comment | EmailEmail | PrintPrint

Enhanced Description of Ontology-like Thing

The following is an enhanced description of an ontology-like thing that is approachable to business professionals.  This definition is inspired and synthesized from the basic textbook definition of an ontology provided in Ontology Engineering by Elisa Kendall and Deborah McGuinness; Michael Uschold's insightful description of an ontology-like things in his presentation Ontologies and Semantics for Industry; and Shawn Riley's description of an ontology's common components in Good Old-Fashioned Expert Systems (With or Without Machine Learning).

An ontology or ontology-like thing is a model that specifies a rich and flexible description of the important

  • terms: (terminology, concepts, nomenclature; primative terms, functional terms);
  • relations: (relationships among and between concepts and individuals; is-a relations, has-a relations; other proerties);
  • assertions: (sentences distinguishing concepts, refining definitions and relationships; axioms, theorems, restrictions); and
  • world view: (reasoning assumptions, identity assumptions)

relevant to a particular domain or area of interest, which generally allows for some certain specific variability, and as consciously unambiguously and completely as is necessary and practical in order to achieve a specific goal or objective or a range of goals/objectives.  An ontology-like thing enables a community to agree on important common terms for capturing meaning or representing a shared understanding of and knowledge in some domain where flexibility/variability is necessary.

And so, the reason for creating an "ontology-like thing" is to make the meaning of a set of terms, relations, and assertions explicit; so that both humans and machines can have a common understanding of what those terms, relations, and assertions mean.  An "instance" or "sets of facts" (a.k.a. individuals) can then be evaluated as being consistent with or inconsistent from some defined ontology-like thing created by that community.

The level of accuracy, precision, fidelity, and resolution expressively encoded within some ontology-like thing depends on the application or applications being created that leverage that ontology-like thing.

An ontological commitment is an agreement by the stakeholders of a community to use some ontology-like thing in a manner that is consistent with the theory of how some domain operates represented by the ontology-like thing.  The commitment is made in order to achieve some specific goal or goals established by the stakeholders in a community sharing the ontology-like thing.

The ontology-like thing and instances (values) created per that ontology-like thing form a sharable conceptualization or logical system that can be tested and proven to be:

  • Consistent (no assertion of the system contradict another assertion)
  • Valid (no false inference from a true premise is possible)
  • Complete (if an assertion is true, then it can be proven; i.e. all assertions exists in the system)
  • Sound (if any assertion is a theorem of the system; then the theorem is true)
  • Fully expressed (if an important term exists in the real world; then the term can be represented within the system)

Every word used to describe an ontology-like thing is there for a reason.  The term "flexible" is included to indicate that we are not talking about a form.  The logical systems we are concerned with have a certain amount of variability and alternatives are allowed, and so the system needs to be extensible.

The system needs to be predictable, reliable, and safe; free from catastrophic failures which would cause undesirable instability.

As pointed out in the ontology spectrum; a dictionary, a thesaurus, a taxonomy, an ontology, and a logical theory are all different types of ontologies.  All types are useful, but what you are trying to get out of the system needs to be matched to what you put into the ontology-like thing.  If you leave one assertion out, errors could creep into the logical system.

There are all sorts of other things that provide the same sorts of functionality as ontology-like things. Many times terms used are different, definitions are somewhat different, what is trying to be achieved is different.  These differences tend to cause confusion and complexity. But the differences tend to be small and the similarities more significant.

Fads, trends, misinformation, politics, and arbitrary preferences all tend to cause distractions from the real choices that need to be made.

The real focus should be on the fact that artificial intelligence applications are brought to life by the metadata provided by ontology-like things.

In particular, high quality curated metadata will supercharge these sorts of applications. Some people say that data is the new oil.   In fact, the Economist declares this in the article, "The world’s most valuable resource is no longer oil, but data."  But others point out, I think correctly, that "If data is the new oil, then metadata is the new gold."

If you read this article, Data Curation: Weaving Raw Data Into Business Gold (Part 1) , the author uses crude oil, refined gasoline, and refined racing fuel as a metaphor to explain the value of metadata. Metadata is simply data about data.  An ontology-like thing is machine-readable metadata.  Curated metadata provided in an ontology-like thing provides the racing fuel used by artificial intelligence applications.

Curated metadata provides that is necessary to make artificial intelligence to work, to supercharge AI.  But creating that metadata takes a lot of work. Machine learning is not a viable short cut. Use short cuts and your AI foundation will be a fragile house of cards.  While machine learning is very, very useful; it is most valuable when it supplements ontology-based things created by humans. Machine learning will never be able to create the initial ontology-like thing.

So, there are two major techniques for implementing artificial intelligence:

  • Logic and rules-based approach (expert systems): Representing processes or systems using logical rules. Uses deductive reasoning.
  • Pattern-based approach (machine learning): Algorithms find patterns in data and infer rules on their own. Uses inductive reasoning; probability.

You can combine both approaches and create a third approach which is a hybrid of both approaches. But you need to use the right tool for the job.

The two letters "AI" appearing on the cover of the Journal of Accountancy helps one recognize the significance of artificial intelligence on accounting, reporting, auditing, and analysis. Ontology-based things is part of that story.  Recognize that XBRL is a syntax for creating ontology-like things for financial reporting.

If there is any confusion in your mind about any of this, I would strongly recommend that you read Computer Empathy. That document provides the background necessary to best absorb this blog post. Successful engineering is about understanding how and why things break or fail.

############

Ontologies, Taxonomies, and Bears -- Oh my!

Fortune: 50% failure rate for enterprises implemneting artificial intelligence

Towards a Theory of Semantic Communication

What is an Ontology?

Forbes: Taxonomies, Ontologies, and Machine Learning: The Future of Knowledge Management

Posted on Friday, July 19, 2019 at 07:33AM by Registered CommenterCharlie in | CommentsPost a Comment | EmailEmail | PrintPrint

Prototype SBRM Represented in XBRL

I have created a prototype SBRM representation using XBRL.  There are many different ways to represent ontology-like things.  I am using XBRL for this representation.  Others will likely use OWL.

My prototype is based on this informal representation of a framework for a financial report.

The the first prototypes created by the OMG working group for SBRM were represented using UML (see page 38 of my Logical Theory Describing a Business Report). What will become very, very clear if the SBRM is represented using both XBRL and OWL is the expressive power of XBRL, the expressive power of OWL, and the real differences between the two representation approaches.  What it is looking like to me is that in order for XBRL to match the expressive power of OWL, XBRL has to re-invent a lot of which already exists in OWL. Whether that makes sense is to be determined.

Here is what I have so far: (note that this is not properly modularized as of yet; this connection approach is optimized for editing of the XBRL taxonomy not using it)

  • Arcroles: First, I defined arcroles using XBRL that will be used to represent relations.
  • Terms: Second, I defined terms. Here are those terms in human-readable form (i.e. this includes relations)
  • Relations: Third, I defined relations. Here is a close approximation of those relations in human readable form.
  • Assertions: It is unclear to me right now if the SBRM will have assertions.  I think that it will not; assertions are defined within an implementation of the SBRM.
  • Instance: The SBRM will not have an instance.
  • Narrative of Conceptulization: Human-readable narrative that describes a financial report in terms a business professional can understand.

Here is an implementation of the SBRM, used to represent a reporting scheme and several reports prepared using that reporting scheme.  I am using the IPSAS reporting scheme.  I am using the method that I documented.  And this is the resulting implementation, the XBRL taxonomy, the XBRL Formulas, all of the other assertions and the XBRL instance of the report.

I will tie all of these pieces together better using the document Illustrating the Benefits of a Best Practice Method for Creating XBRL-based Financial Reports.

For reference purposes, here is a prototype of SBRM represented in OWL (this is a work in progress). You can download and install Protege which will load this.  This is a screen shot of what it looks like thus far in Protege.

That is it for now.  Stay tuned!

Posted on Sunday, July 14, 2019 at 09:03AM by Registered CommenterCharlie in | CommentsPost a Comment | EmailEmail | PrintPrint

Ontology-like Things for Industry

Michael Uschold, Senior Ontology Consultant with Semantic Arts, provides this presentation Ontologies and Semantics for Industry where he explains the benefits of ontologies and ontology-like things to industry.

Ushold points out that there is a plethora of 'ontology-like things'.  I like the term "ontology-like things". That is what the ontology spectrum tries to point out.  Essentially, there are many different ways to express meaning, an ontology is just one of those ways.

The presentation is organized to answer three questions:

  1. What is the difference between an Ontology and a:
  2. When people say things like "Ontologies have unambiguous semantics", what do you think they mean; do you believe them; and why or why not.
  3. How are ontologies and semantics relevant to industry today?

Ontology is defined in the textbook Ontology Engineering  by Elisa Kendall and Deborah McGuinness as follows:

Ontology - a model that specifies a rich description of the

  • terminology, concepts, nomenclature;
  • relationships among and between concepts and individuals; and
  • sentences distinguishing concepts, refining definitions and relationships (constraints, restrictions, regular expressions)

      relevant to a particular domain or area of interest.

I reconcile that definition above to the common components of an ontology that I summarize in the document demystifying ontologies as follows:

  • Terms
    • Simple terms (primitive, atomic)
    • Functional component terms (complex functional terms)
    • Properties (qualities, traits)
  • Relations
    • Type relations (class/type relations, "type-of" or "is-a" or "class-subclass" or "general-special")
    • Functional relations (structural relations, "has-a" or "part-of" or "has-part" or "whole-part")
    • Property attribution (has property)
  • Assertions
    • Restrictions (constraints, limitations)
    • Axioms
    • Rules (theorems)
  • Individuals
    • Instance (facts)

This forms a formal, logical system that is:

  • Consistent (no theorems of the system contridict one another)
  • Valid (no false inference from a true premise is possible)
  • Complete (if an assertion is true, then it can be proven; i.e. all theorem exists in the system)
  • Sound (if any assertion is a theorem of the system; then the theorem is true)
  • Fully expressed (if an important term exists in the real world; then the term can be represented within the system)

Fundamentally, Uschold points out, an ontology is "A way for a community to agree on common terms for capturing meaning or representing knowledge in some domain."  The ontology spectrum helps you understand to what extent you are actually agreeing. (Logic; Formal System)

And so, the reason for creating an "ontology-like thing" is to make the meaning of a set of terms, relations, and assertions explicit, so that both humans and machines can have a common understanding of what those terms, relations, and assertions mean.  "Instances" or "sets of facts" (a.k.a. individuals) can be evaluated as being consistent with or inconsistent with some defined ontology-like thing created by some community.  The level of accuracy, precision, fidelity, and resolution expressively encoded within some ontology-like thing depends on the application or applications being created that leverage that ontology-like thing.

An ontological commitment is an agreement by a community to use some ontology-like thing in a manner that is consistent with the theory of how some domain operates, represented by the ontology-like thing.  The commitment is made in order to achieve some goal established by the community sharing the ontology-like thing.

Ontology-like things for accounting, reporting, auditing, and analysis require high-quality and therefore they require highly expressive ontology-like things.

As Kendall and McGuinness point out, "The foundation for the machine-interpretable aspects of knowledge representation lies in a combination of set theory and formal logic."

Put these pieces together correctly and you can get software applications to perform magic! That is what curated metadata is all about. You need to use the right tool for the job. Financial reports will lead the way.

Might be time to increase your digital maturity.

Posted on Saturday, July 13, 2019 at 07:10AM by Registered CommenterCharlie in | CommentsPost a Comment | EmailEmail | PrintPrint