On Sat, Nov 13, 2021 at 12:14 AM Ben Goertzel <[email protected]> wrote:

> > Sorry, the question was not whether or not AtomSpace is used for each
> project independently.  (thank you for the detailed answer though)  It was
> meant to ask if multiple projects use AtomSpace as a blackboard in order to
> interoperate.
>
>
> They don't at the moment, in practice.
>
> For this to work, one would need a common ontology across the multiple
> projects.   The ontology implicit in the built-in set of Atom types is
> generally too low-level to admit practical unification of multiple
> applied Atomspace projects.
>
> In our AI-DSL project at SNet (which does not use Atomspace at
> present, but probably will in future when we shift it from Idris2 to
> MeTTa as a base language), we are using SUMO as a base ontology for a
> similar purpose
>
> One would then need to e.g. map the ontology used in the current
> OpenCog Bio-Atomspace into SUMO (which would not be so hard)
>
> I have been discussing w/ Josef Urban (who is collaborating with Adam
> Pease the SUMO dude) some projects using neural models to
> automatically map NLP expressions into SUMO interpretations.


Nifty, Adam in 1999 employed me for a couple year to write the first SUMO
theorem prover which can be found here:
https://github.com/logicmoo/sigma_ace/tree/master/engine (later on it was
replaced by Vampire-KIF)

I use SUMO+CycL  as the output of my NLP code as well
 (can see the output here)
https://docs.google.com/document/d/1IELBApotKwFxTvsUsEgU5nPqofuHVsyuDzxTBw4ca_A/edit



> This
> would be a narrow AI project obviously not AGI, but could be useful as
> a tool supporting some sorts of AGI dev...
>


I see nothing wrong with that idea, in fact I am using my NLU->SUMO
translator to help produce narrative stories to provide the initial content
for a virtual world in which an AGI can train in.   Also since my world is
maintained as a SUMO/CycL KB  and my translator works both ways
(English->Sumo and Sumo->English) It makes for making the world and rules
that make it up easier to debug.




>
> -- Ben
>
> --
> You received this message because you are subscribed to the Google Groups
> "opencog" group.
> To unsubscribe from this group and stop receiving emails from it, send an
> email to [email protected].
> To view this discussion on the web visit
> https://groups.google.com/d/msgid/opencog/CACYTDBeuyYE0YFQ5e%2B-rsmwe2r%3Da18pe0uJTgvv-wKUT7Qo3Sw%40mail.gmail.com
> .
>

-- 
You received this message because you are subscribed to the Google Groups 
"opencog" group.
To unsubscribe from this group and stop receiving emails from it, send an email 
to [email protected].
To view this discussion on the web visit 
https://groups.google.com/d/msgid/opencog/CAER3M5k1zdtft3%2BRN%3Dqx%3D0hCfJy8yqB-g7Y84K5x344AejXUFg%40mail.gmail.com.

Reply via email to