Project History


Professor Doug Lenat

The Cyc project, a long-term quest to develop a true artificial intelligence, was founded in 1984 by Dr. Douglas Lenat as a lead project in the Microelectronics and Computer Technology Corporation (MCC). The work in Expert Systems that took place during the 1970's and early 80's produced some impressive results in areas including medical diagnosis, oil exploration, and computer configuration, but the resulting systems were all dangeroupsly brittle. 

MCC, Austin, TX

That is, they each had an extremely shallow understanding of the domain they purported to model. Medical systems understood almost nothing about humans or diseases; oil exploration systems didn't know that oil was a fluid or that it could be used as fuel; computer configuration systems didn't understand what computers were other than being abstract systems of components. While these surface models were sufficient to perform their intended tasks, this lack of a real-world understanding of their domain meant that they were not easily extensible, the models were not very reusable, and the systems were subject to failures or mistakes that no ten-year-old child would ever make. Some leaders in the AI community at the time recognized that if software was every to be "intelligent", it would need to have a much richer understanding of the world. Creating a system with a sufficiently broad and deep understanding, as well as the infrastructure to represent and reason efficiently with such knowledge, was a daunting task. Recognizing the significance of this work, and the need for someone to achieve it, Dr. Lenat began creating a corpus of machine-understandable knowledge and an inference engine to reason over it. The initial phases of this work was conducted as a long-term, high-risk, high-payoff project at MCC.

After about a decade of infrastructure development and knowledge base construction, the project was spun out of MCC and into a private company. In 1994, Cycorp was founded to further develop, commercialize, and apply the Cyc technology. With support from a blend of commercial and government clients, including both R&D and application development, the Cyc technology has continued to evolve.

On-going R&D projects focus on using Cyc to provide deep understanding of text, richly interface with structured knowledge stores, efficiently reason over vast amounts of knowledge and data, and automated machine learning. Cycorp provides commercial solutions to Fortune 100 companies in industries including energy, healthcare, finance, and education and offers a semantics solution development and deployment platform for those who wish to develop their own knowledge-based applications. To foster the growth of semantic modeling and machine reasoning, Cycorp offers a no-cost license to its semantic technologies development toolkit to the research community. Currently over 350 research licenses have been granted to universities as well as government and commercial research laboratories. Additionally, Cycorp has placed the core Cyc ontology into the public domain and hundreds of thousands of copies have been downloaded.