Fantastic, thank you! This is an information theoretic measure of surprise right?
Is it related the the Jenson-Shannon Divergence? Is there a paper on it? Johannes On Wednesday, February 20, 2019 at 5:41:20 AM UTC, Nil wrote: > > Hey, > > I have ported Shujing's I-Surprisingness to the new pattern miner and > I'm generalizing it for patterns with partial linkage between clauses. > The problem with Shujing's code is that it only works with fully > connected clauses like > > And > P(X,Y) > Q(X,Y) > > but not with partially connected ones like > > And > P(X,Y) > Q(Y,Z) > > usually rating these as more surprising than they are. I except to have > something ready by the end of the week. It won't be the grand theory of > surprisingness we'll eventually have to converge to, but it should be > good enough to filter out a big class of obviously unsurprising patterns. > > Nil > > On 2/19/19 10:32 PM, Johannes Castner wrote: > > Hi Nil, > > > > How is the interesting pattern-miner doing these days? I'm getting > > closer and closer to the point where I'd love to run it on some things. > > > > Hope you're doing well! > > > > Johannes > > > > On Thursday, February 7, 2019 at 7:33:34 AM UTC, Nil wrote: > > > > Hi Linas, > > > > On 2/6/19 10:53 PM, Linas Vepstas wrote: > > > What's the bug? > > > > The pattern matcher fails to consider some candidates if they have > > been added before some rule-engine call. I wonder if it is another > > manifestation of > > > > https://github.com/opencog/atomspace/issues/1921 > > <https://github.com/opencog/atomspace/issues/1921> > > > > which, if so, means it goes even deeper than the pattern matcher. > > > > > I encourage open discussion first; it can sometimes help > > > avoid wasting time on dead-ends. > > > > I totally agree, but at this point I'm still trying to narrow the > > reproducability of the problem down to a minimal number of steps > > before creating an issue. > > > > Nil > > > > > > > > -- linas > > > > > > -- > > > cassette tapes - analog TV - film cameras - you > > > > > > -- > > > 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] <javascript:> > > > <mailto:[email protected] <javascript:>>. > > > To post to this group, send email to [email protected] > > <javascript:> > > > <mailto:[email protected] <javascript:>>. > > > Visit this group at https://groups.google.com/group/opencog > > <https://groups.google.com/group/opencog>. > > > To view this discussion on the web visit > > > > > > https://groups.google.com/d/msgid/opencog/CAHrUA36WY5_wXXvnuzb6arMORGgOpicteubLNfzCmBjnRX%3DaOw%40mail.gmail.com > > > < > https://groups.google.com/d/msgid/opencog/CAHrUA36WY5_wXXvnuzb6arMORGgOpicteubLNfzCmBjnRX%3DaOw%40mail.gmail.com> > > > > > > > > > < > https://groups.google.com/d/msgid/opencog/CAHrUA36WY5_wXXvnuzb6arMORGgOpicteubLNfzCmBjnRX%3DaOw%40mail.gmail.com?utm_medium=email&utm_source=footer > > > < > https://groups.google.com/d/msgid/opencog/CAHrUA36WY5_wXXvnuzb6arMORGgOpicteubLNfzCmBjnRX%3DaOw%40mail.gmail.com?utm_medium=email&utm_source=footer>>. > > > > > > > For more options, visit https://groups.google.com/d/optout > > <https://groups.google.com/d/optout>. > > > > -- > > 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] <javascript:> > > <mailto:[email protected] <javascript:>>. > > To post to this group, send email to [email protected] > <javascript:> > > <mailto:[email protected] <javascript:>>. > > 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/e4aa2518-e6cc-4ee6-94e4-45c753961a80%40googlegroups.com > > > < > https://groups.google.com/d/msgid/opencog/e4aa2518-e6cc-4ee6-94e4-45c753961a80%40googlegroups.com?utm_medium=email&utm_source=footer>. > > > > For more options, visit https://groups.google.com/d/optout. > -- 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/f2033ea5-6a48-4121-a8e8-61ffa47de928%40googlegroups.com. For more options, visit https://groups.google.com/d/optout.
