Le mar. 9 nov. 2021 à 18:18, Linas Vepstas <[email protected]> a écrit : > > On Tue, Nov 9, 2021 at 3:01 AM Amirouche Boubekki > <[email protected]> wrote: > > > > There is one major missing piece, I do not know how to word it properly. > > Let's try: Linas' algorithm does build a model in a way that is > > unsupervised that can predict the linguistic structure of a sentence. With > > that model, Every word part a sentence can be attached to half a directed > > link, in the spirit of Link Grammar, then when the sentence make sense, > > there is complete linkage of the sentence with full directed links. Where > > does a link label come from? > > The label is whatever you want it to be. The algo doesn't care. > > I have no idea why you think this is a "missing piece", never mind it > being "major".
It is a missing piece to be able to tell whether the linkage make sense or not. #explainableai > > BTW, I'm convinced the algo works for audio and video, too. I'm > convinced that it also works for higher layers of abstraction, too. Yeah, but then how do you build e.g. semantic frames or pragmatics from the linkage? -- 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/CAL7_Mo-1wd76-Vv_dkcO5tCo%2B7Nn7P0-S5tg6OV4jeVc_mDrjw%40mail.gmail.com.
