> The R2L code does have some fundamental design flaws: the rules are > hand-coded, there are about 60 or so of them, and we really need twice as > many, but even if we had more, they most they can deal with is relatively > unambiguous factual English sentences. What we really need is a way to > automatically learn such rules.
What Ruiting is working on now is a way to automatically learn those rules, via pattern mining on an English/Lojban parallel corpus... Other was of learning such rules are also possible... But for simple prototype-y examples the R2L rulebase is usable and does something, yeah... 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 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/CACYTDBfiN3cOYi7KOJ_3N1GFnhERE4SP1uUdqNwKKAVw%2B%3D_jfA%40mail.gmail.com. For more options, visit https://groups.google.com/d/optout.
