The difference
  • 1
    Symbolic + Neural network technologies
    This is a new Symbolic Paradigm, which combines the advantages of both the symbolic and neural network technologies. No one else has the structures we have developed.
  • 2
    Natural language reasoning
    Not just "Processing," but "Reasoning". This involves understanding the text, and not just simple pattern matching and word counting to determine the meaning of things.
  • 3
    Linear, qualitative 
    We don't simply tag a word with an arbitrary number that determines the relevance of some data. We can follow a train of thought that makes sense of an argument, in terms of understandable logic.
  • 4
    When you work with neural networks, everything is distilled into a set of numbers, which tell us nothing about what is actually going on. Our new symbolic paradigm can be traced end-to-end, and may be surgically corrected.
  • 5
    Augmented topological network
    We have a new structure, a network, but augmented by what its connectivity entails, and the arrangement of its constituents. Ontologically topological, in that we use the meanings and not specific terms to operate.
  • 6
    Human-understandable logic
    The logical process that our new intelligence follows we can trace, step by step, and understand what is going on "under the covers". Any false step will show itself in its functioning.
  • 7
    It is a universal logic that we can model. Anywhere that human logic can apply, we can go there, too. We are not limited to any narrow discipline: as long as we can talk about it, we can understand it.
  • 8
    Any language
    As long as humans can make sense of it, our intelligence can, too. We are dealing with the meanings of things, not their specific expressions, though we know about those too.
  • 9
    Don’t need tons of data
    We can learn from example. We can generalize. You can tell us something one time, and if it makes sense, that's it: we know it forever. No need to hammer something down again and again to match some pattern.