The world’s economy and society is increasingly knowledge-based.
Our vision is to enable automated systems to radically better accumulate and harness knowledge. This includes to:
- communicate better with people and other systems: at the levels of deep meaning and reasoning
- learn by better harnessing humanity’s accumulated storehouse of distilled knowledge; this is far more powerful than raw learning from observation — i.e., data mining
- scale up intelligent systems to very large, widely-authored knowledge bases that converse, answer questions, and proactively supply useful information
Relational databases were the first successful major semantic technology. They revolutionized enterprise computing. Our approach is a next generation database system that we believe will fundamentally change the knowledge representation playing field once again in semantic technology — as well as in related large swaths of AI, data analytics, and software engineering.
This has the potential to create an even more powerful virtuous circle in accumulating and harnessing knowledge, both in the enterprise space and beyond.
We introduce the concept of knowledge representation and reasoning (KRR) in human-machine logic – “humagic”™ for short. Broadly speaking, humagic is the hybrid of human thought and machine action, at once easily understandable to people and capable of deep reasoning by computers. We envision a continuum from human ideas, to verbal articulation, to written articulation to encoded ‘smart’ sentences, to highly expressive and flexible knowledge bases in Ergo that are capable of analysis, alerts, decision making, question answering, explanations, and more.
Operationalized in Rulelog, and Coherent’s implementation, Ergo Suite, humagic knowledge includes sentences in both natural language and logical syntax, closely linked to each other as part of overall knowledge management. These sentences may be assertions, queries, or conclusions. For example, a complex English sentence about a financial regulation in a legal document may be linked to an executable complex logical sentence assertion that encodes that English sentence. And a complex logical sentence conclusion may be linked to an English sentence that presents the logical sentence for an end user. The links themselves are humagic knowledge too; sometimes as assertions and sometimes as conclusions, easily understandable to both machine and human users. Humagic knowledge includes the full range of collaborative development and evolution of knowledge, as knowledge is specified, made operational, debugged, and extended.