>What if we decompose all semantics into unambiguous ​units?  Then each
vector has only 1 sense.

That is an interesting idea but is it practical? All the good AGI
techniques work for some cases but don't work for other cases, and it seems
that the situations that they don't work well in are so widely distributed
that you can't really get any of the methods to gain enough traction to get
going. (I am only guessing at this from reports of methods that other
people have tried.) I don't know if Watson is evidence that this might be
solved by computing power or not.

I mentioned an idea I had a few weeks ago where an artificial language
could be based on primitives (and combinations) which might be ambiguous
but if artificial conversation was constrained in such a way so that the
ambiguities would be reduced then the language could be feasible. An idea
like that might make your goal more feasible, but it would be difficult to
get something like that to work outside of a simple pilot program.
Jim Bromer

*If you can solve a problem by avoiding it then your attitude may be part
of the problem.*


On Thu, Jun 19, 2014 at 4:48 PM, YKY (Yan King Yin, 甄景贤) <
[email protected]> wrote:

> On Thu, Jun 19, 2014 at 3:16 AM, Jim Bromer via AGI <[email protected]>
> wrote:
>
>> If all active words (verbs) had only one or two directions then a
>> directed vector could be used. However, part of the value of words is that
>> you can assign a variety of possible relations to them and a single
>> standard will not work. Suddenly the space and the mathematics has to be
>> warped to cover the variations. Vector Space and Semantics is just not a
>> good match.
>> Jim Bromer
>>
>> *If you can solve a problem by avoiding it then your attitude may be part
>> of the problem.*
>>
>
> ​
> What if we decompose all semantics into unambiguous ​units?  Then each
> vector has only 1 sense.
>
>



-------------------------------------------
AGI
Archives: https://www.listbox.com/member/archive/303/=now
RSS Feed: https://www.listbox.com/member/archive/rss/303/21088071-f452e424
Modify Your Subscription: 
https://www.listbox.com/member/?member_id=21088071&id_secret=21088071-58d57657
Powered by Listbox: http://www.listbox.com

Reply via email to