Learn

  • a "chair" is both a collection of all "facts" one might know about chairs, as well as an English word participating in the complex grammatical network of the English language. The word "chair" is not so much a "symbol" as it is a bridge between these complex networks... the goal of this project is to not only learn what a "chair" is from environmental stimulous, and not only learn how to use the word "chair" in a grammatical sentence, but also to learn the rules of logic, and the rules of common sense reasoning, out of thing air.
  • The way in which regions of graphs look locally similar can be described by a grammar. The way they connect is a syntax.
  • If you are writing text, what you are "actually doing" is taking the network of interconnected facts/ideas, and serializing them into a sequence of words. A sequential, time-like order, one word after another. You serialize the graph of ideas.
  • In a text, there is both a small-scale structure, and a large scale structure. The small scale structure consists of grammatically-correct sentences written in some natural language. The large-scale structure is determined by the type of the text: is it a short story? A dissertation? An owner's manual? The large scale flows from sentence to sentence, paragraph to paragraph, as repeating ideas, characters, topics, themes reoccur throughout the text
  • all work must be done in a single, common framework, that makes it easy to collect statistical information from a large variety of sources. A place where the statistical analysis is easy. A place where inference can be automated. A place where rules can be defineed, generically, and lanuched, generically. That single location is, for me, the AtomSpace