🚀 go-pugleaf

Seth Russell wrote:
> ...
> Does anyone know of any research that attempts to feed the second
> approach with the output of the first approach ?   Computational
> linguistics can form a pretty accurate respresentation of a stimulus -
> for example the parse tree of a question.  Has anyone fed that to a
> nueral net, the nueral net responding with a parse tree of the answer ?

I don't know who is working on that right now, but it is clear that this
one of the most likely ways we will take the next step in the
utility/power of AI systems.  Efforts like Cyc and LOOM are probably
going to yeild useful reasoning systems but will eventually run into
trouble due to the hand crafted nature of their structure.  By the time
that happens we can expect some high powered hardware that does its own
job of classification in the neural style.  Flip the switch to train the
new system from the old and you're ready to tackle a whole new level of
problems.

jim
-----------------------------------------------------------------------
James P. White                        Netscape DevEdge Champion for IFC
Director of Technology Adventure Online Gaming http://www.gameworld.com
Developers of Gameworld -- Live Action Role-Playing and Strategic Games
jim@aognet.net            Pagesmiths' home is http://www.pagesmiths.com