On Fri, Apr 28, 2017 at 10:05 AM, Ben Goertzel <[email protected]> wrote:
> Hi, > > > Just got out of surgery for my broken leg; this email attempts to prove > that > > the general anesthesia didn't kill too many brain cells. > > Oh shit... so that little ankle-twist you got in Lalibela was actually > a broken bone??!! ;o that would explain why it was taking so f**king > long to heal ... ;/ > Yeah. and I think it was even healing correctly, until one day when it was actually getting in pretty good shape, I was hurrying and made it worse, because it deteriorated after that. > > > Yeah, cool! We are definitely getting somewhere. These results are > all with an overly small corpus (as well as being the first stage of > the iterative process) so some dopey mistakes are to be expected... > Well, but this is all old ground. I was seeing this years ago, and besides, its what all the old MST parsers from a decade ago were all about. Its also why they were abandoned: no one figured out how to make them better, by extracting link types. So the next step will be a first. > > Yes. Ruiting is ready to start playing with clustering approaches, > as soon as you give her an Atomspace of mst-parsed sentences that you > think is a passable corpus for clustering experimentation... > OK, "real soon now". There's always a few hiccups. The mst parser, being scheme code that calls the atomspace, is at least a magnitude slower than the LG parser (which has gotten really fast). Note, however, that the it is possible to export all of the MST data into LG, and use LG to perform MST. This has been in planning for a while, and the ground has been cleared for this, but remains to be implemented. --linas -- 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 post to this group, send email to [email protected]. Visit this group at https://groups.google.com/group/opencog. To view this discussion on the web visit https://groups.google.com/d/msgid/opencog/CAHrUA34T7POKZR%3Dptm9bvhKU3gsYOXw1vwTa9P1Q2-01_pgaEw%40mail.gmail.com. For more options, visit https://groups.google.com/d/optout.
