Cognonto today announced the release of version 1.20 of KBpedia, its knowledge structure that integrates six major knowledge bases (Wikipedia, Wikidata, OpenCyc, GeoNames, DBpedia and UMBEL) and 20 subsidiary ones under the KBpedia Knowledge Ontology (KKO). KBpedia’s explicit purpose is to provide a foundation for knowledge-based artificial intelligence (KBAI) by supporting the (nearly) automatic creation of training corpuses and positive and negative training sets and feature sets for deep, unsupervised and supervised machine learning.
The changes in this new release are solely related to KKO, the knowledge graph portion of KBpedia. There are two major drivers for this update to the KBpedia upper ontology. First, internal development efforts are now focusing on the modeling of predicates and time and action. This effort affects the definitions, splits and boundaries between attributes, relations, events and activities. Revisions in this area have been derived from a much closer reading of Charles Sanders Peirce‘s writings, based on our view that CSP has the most logical and sophisticated understanding of knowledge representation yet expressed. Second, where appropriate, we have relied on Peircean terminology to capture specific concepts. We are doing this to make KKO more amenable to review by Peircean scholars. At the same time, we have tried to reduce the use of obscure or difficult Peircean terms where they might be a barrier to understanding.
These changes solely affect two of the three main branches in KKO. The most affected branch is Monads, the branch representing Firstness (see below), reflecting the basic concepts or building blocks used in KKO. The Particulars branch, which captures the representation of individuals or instances, also was modified to capture those changes in the Monads branch. The Generals branch, the main portion for classes and types, was not affected by these changes.
The resulting KKO upper structure now has this form with about 165 key concepts, all organized according to Peirce’s universal categories of Firstness (1ns), Secondness (2ns) and Thirdness (3ns) (I earlier provided a broad overview for the basis of this triadic design):
|level 1||level 2||level 3||level 4||level 5||level 6||level 7|
|Generals [3ns] (== SuperTypes)|
It is useful to re-cap the three constituents of Peirce’s trichotomy, what he called simply the Three Categories, or the universal categories, as follows:
- Firstness [1ns] — these are possibilities or potentials, the basic forces or qualities that combine together or interact in various ways to enable the real things we perceive in the world, such as matter, life and ideas. These are the unrealized building blocks, or elements, the essences or attributes or possible juxtapositions. They are not divisible, what Peirce called indecomposables, since they are integral qualities or ideas in themselves;
- Secondness [2ns] — these are the particular realized things or concepts in the world, what we can perceive, point to and describe. A particular is also known as an event, entity, instance or individual; and
- Thirdness [3ns] — these are the laws, habits, regularities and continuities that may be generalized from particulars. All generals — what are also known as classes, kinds or types — belong to this category. The process of finding and deriving these generalities also leads to new insights or emergent properties, which continue to fuel knowledge discovery. Insights arising from Thirdness enable us to further explore and understand things, and is a driving force for further categorization.
Note that the three main branches and most of the sub-branches to KKO conform to this triadic structure. The basis for this structure was discussed in an earlier article.
In future posts I will delve and explain further each of the main branches of KKO. It is also likely the changes and refinements to this upper structure may continue for some time. Cognonto has open sourced KKO both for use by others and as a means for Peircean scholars and students to make critical commentary and suggestions. Because of this desire for review, we have also annotated the KKO structure more extensively in this release, including references to specific passages from Peirce’s writings.
Remember, KBpedia and KKO are the first complete attempt to capture Charles S. Peirce’s views of the logical organization of knowledge and the theory of signs into a working computer ontology (knowledge graph). As with Peirce’s views of ‘truth‘ as a limit function that can be approached but never fully attained, we will continue to strive to improve our understanding of how best to model knowledge for artificial intelligence purposes. The good news is we are already realizing significant KBAI benefits from KBpedia in its current form. We expect those benefits to continue to grow with further refinements to KKO and its typologies.
The open source KBpedia Knowledge Ontology (KKO) may be downloaded and inspected from here. We welcome any and all critical commentary.