Cognonto, a new start-up in knowledge-based artificial intelligence (KBAI), announced today the dual release of its Cognonto Platform and KBpedia, a computable knowledge structure to automate much of the effort needed for machine learning. KBpedia leverages six large-scale knowledge bases — Wikipedia, Wikidata, GeoNames, OpenCyc, DBpedia and UMBEL — into a single structure expressly designed to support artificial intelligence (AI) within enterprises.
“Many of the AI advances in recent years, such as question answering on smart phones or systems that beat human contestants in Jeopardy, are built around Web knowledge bases like Wikipedia,” said Michael Bergman, a co-founder of Cognonto. “But these are one-off systems that only the largest tech firms or research outfits can afford,” he said. “The idea behind Cognonto is to democratize this process such that any enterprise can afford to train their own machine learners or gain the advantages of knowledge-based artificial intelligence.”
KBpedia combines the hundreds of thousands of concepts and 20 million entities in its source knowledge bases in a structure that separately captures entities, attributes, relations and topics, the types by which they are categorized, and the connections between them. One innovation of the system, according to Bergman, is the schema, or “knowledge graph,” that organizes KBpedia according to the logic of Charles Sanders Peirce, a noted 19th century American mathematician, philosopher and polymath.
Cognonto offers a variety of enterprise services via either dedicated SaaS (software as a service) or deployment of the Cognonto Platform. “While KBpedia is useful in a standalone mode, most of our clients want to integrate with their own sources and data,” said Frédérick Giasson, Cognonto’s other co-founder and CTO. He noted the Cognonto Platform also leverages KBpedia to map and integrate enterprise content to tailor the machine learning. Giasson said typical uses include fine-grained entity extraction, categorizing content, and enterprise information integration.
Cognonto has been under active development for more than two years, though the genesis of the venture reaches back nearly six years. Its principal founders are noted leaders, innovators and bloggers in the semantic technology space, with most recent efforts focused on KBAI.
Cognonto (a portmanteau of ‘cognition’ and ‘ontology’) exploits large-scale knowledge bases and semantic technologies for machine learning, data interoperability and mapping, and fact and entity extraction and tagging. Cognonto LLC is privately held and is based in Coralville, Iowa, USA, with offices in Québec City, Québec, Canada.