Hi Nil, We need to get back to Shujing on all these issues ASAP so she can proceed.
I will think/look carefully regarding the Interestingness Measure issue, as the probabilistic surprisingness aspect of this is something I have thought about a lot... If you can look at the other issues carefully today or tomorrow, that will be valuable... thanks a lot Shujing: About the definition of "pattern gram", OK I see how the code works now. But it seems to me that counting the number of "root links" is a bit arbitrary, isn't it? The number of Atoms in a pattern altogether is meaningful, and the number of variables in the pattern is meaningful (and I guess how "loose" are the type restrictions on these variables, is also meaningful). But the number of root links doesn't mean much, as a root link can supervene over a huge or tiny tree of Atoms, right? Hypergraphs are complicated ;p ... ben On Mon, Jun 19, 2017 at 6:10 AM, Shujing Ke <[email protected]> wrote: > Anyone looked at the sample file yet? > Because each Link is big, my machine cannot finish mining all the patterns > from the whole corpus with current setting. It would be nice if anyone can > give more ideas about what kinds of patterns should be kept and what kinds > of patterns can be filtered out in the sample file. > > Thanks, > Shujing > > On Thu, Jun 15, 2017 at 12:19 AM, Shujing Ke <[email protected]> wrote: >> >> Pattern gram in the current implementation is about the numbers of root >> Links one pattern contains. A 1-gram pattern can contain many many >> variablenodes if there are many many Links nested in a root Link, in this >> case, the ExecutionOutputLinks . >> >> On Wed, Jun 14, 2017 at 4:01 PM, Ben Goertzel <[email protected]> wrote: >>> >>> Hi, >>> >>> I'm busy tonight and tomorrow but can look at this Friday, unless Nil >>> has studied it first ;) >>> >>> About >>> >>> > GroundedSchemaNode and TypeNode are not considered to become >>> > VariableNodes. >>> > Seems it doesn't make a lot of sense to make them into variables. >>> >>> well it does make sense to make these into variables, but it may not >>> be urgent... >>> >>> In dependent type theory one has lots of type-valued variables, it's >>> actually critical to programming in languages like Agda that are based >>> on dependent types and Curry-Howard correspondence, etc. >>> >>> About pattern gram, I'm a little confused. We do want patterns >>> involving more than one PatternVariableNode, in mining patterns from >>> the PLN inference histories... don't we? >>> >>> ben >>> >>> >>> -- >>> Ben Goertzel, PhD >>> http://goertzel.org >>> >>> "I am God! I am nothing, I'm play, I am freedom, I am life. I am the >>> boundary, I am the peak." -- Alexander Scriabin >> >> > > -- > 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/CALpD4-%2B2e0YrjhxTdzspaSOCZmHSwd2RMRw2KF0PGsP1epYwxg%40mail.gmail.com. > > For more options, visit https://groups.google.com/d/optout. -- Ben Goertzel, PhD http://goertzel.org "I am God! I am nothing, I'm play, I am freedom, I am life. I am the boundary, I am the peak." -- Alexander Scriabin -- 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/CACYTDBevM_AigNPpUqWAttf-fvsw0M%2BNzhjNzUyntXP_P_zZhA%40mail.gmail.com. For more options, visit https://groups.google.com/d/optout.
