Genesis of the Venture



CORALVILLE, IA, October 16, 2014 -- For more than six years, we have been actively studying Wikipedia and its potential role in the semantic Web and for knowledge-based artificial intelligence (KBAI). Our 'SWEETpedia' compilation of hundreds of technical articles leveraging the structure and content of Wikipedia has been a go-to resource on these questions since early 2008.

Despite its unprecedented size, scale and scope (4.5 million articles in the English version; 245 current language versions), the results of this external research supports the view that Wikipedia has an ineffective category structure and is not adequate as is to support computer reasoning. Though highly useful for specific question lookups, Wikipedia, without modification or re-expression, has not been meeting its broader AI and related potential.

Our efforts to map the Wikipedia content to our own UMBEL reference structure began in 2010. Through these mappings, we began to get client engagements in entity extraction and in training machine learners for categorization and tagging. Efforts like DBpedia were also showing the rich content structure underlying Wikipedia. As our mappings increased, we began to realize we could automate the most costly portion of these supervised machine learning projects: the creating of positive and negative training sets.

This 'Aha!' moment in September caused us to see that Wikipedia and related projects could produce a feature-rich, data-driven structure that we could express using a logical and coherent knowledge graph (also known as an 'ontology'). The growing success of Wikidata for instance data (records) was a complimentary spark for our 'Aha!' moment. The integration of our concept schema leveraging Wikipedia and Cyc -- with huge numbers of instance data becoming available on the Web -- would ensure we could integrate multiple data sources into a computable whole.

Armed with this powerful venture focus, we then began searching for a name and a logo. We eventually embraced 'Cognonto' as our venture name (a portmanteau of cognition and ontology), and secured Web site names. Once the name was decided, we conducted an open competition to produce the Cognonto logo. We chose our new logo today. We love it; it will be hard to keep it a secret until we go live with the venture!

During the open logo competition, we began the essential process of writing our business plan. I have not met a plan yet that was not obsolete once it was written. But I have also not met a venture that has succeeded without the effort to produce such a plan, since it forces one to consider all of the various factors required. Development of our business plan (and the processes for updating it) is now well underway.

Profits and available development time, which for us unfortunately act in opposition to one another, will be the governing factors determining how long it will take to get the Cognonto venture ready for release. The real news is that our commitment to the Cognonto venture has now been made. It is now time to execute!

Note: This release was written for internal documentation purposes and was not publicly released until the inauguration of the formal Cognonto venture.