this is just a copy a paste i just want to put it on the internet, its
this class is headed off to coding human senses.
WHAT YOU GOTTA DO
* i want you to be able to predict the future of a data stream. i need
to know if the machine predicted it, or it was anomolous to it.
the more you score prediction hits the better a grade you get.
* form the most common groups of bits, and common sequences of groups
* then group these groups and form sequences of these, this will let
you predict further and compress the information of the larger entities.
and one last thing to add for the class is
* the better you form the groups, the more closer they are to an entity
in the data, or a word, or SYMANTIC.
* grouping the data temporally and spacially is important, however one
decides to say something is space and something is time is
COMPLETELY UP TO YOU. because that is data dependant, but i think
spacial techniques work for time and time techniques work for space
and its actually interchangable.
* making the sequences vary in distance, and the words vary in size is
VERY IMPORTANT TO GETTING IT TO BE A REAL SENSE. you must
“carve” the most COMMON words you can, if theres a little too much in
the word, or a little too less, then its NOT A GOOD CUT OF THE BITS!
* if your doing pictures, your literally chipping off the objects from
the background, if you code your bit carver right!
so imagine looking at the data ien different segment sizes, and your
picking off segments youve seen most often.
if its reoccuring its not by accident its reoccuring (in the almighty
words of j. hawkins.)
so what we want is the damn FINEST REOCCURING GROUPS! then!! we will
not be just chopping up data randomly well be finding very
IMPORTANT BITS OF IT!!!
see, its coming clear now.
its pretty good. and it has fuzzy error correction.
so wed wind up with the most common sentences, nothing more.
so let me think… how do rules of grammar get into it?
so its got these common sentences, like maybe 2 or 3 words at a time,
like “hi man” or “you got it” catch phrases that occured often
i never thought of how simply it applys to language.
remember, it working on moving pictures too makes it look even better.
but yeh, im thinking now, its not really got rules of grammar in it,
like i jumped to.
well its time to think!!!!!!
i never thought of it so simply before.
it like will store the personality of the speaker inside it.
it could score your writing for commonalities, and score all your words
for how often you say them!!! hehe
well, starting from letters is pointless for it anyway, because you have
the symbols already, it actually getting the “catch phrases” of
the motion video is more exciting, then we have no method of gaining
words from the data except THIS!
see, grok!!! is showing off phrases, but its not set in a heirarchy just
yet, its even better then, its showing off the catch phrases of
a machine in action.
but you can see… now we are supposed to be able to add motor to this
collection of catch phrases!?>!?!?!?!?!?!??
well the funny thing will be able to spit out its version of what went
into it, in its funny catch phrasey way.
and thats what magnus was primarily interested in, as a form of
entertainment perhaps, he was a toy maker after all.
you should be able to, with storing ENOUGH INFORMATION, with the SET
COMPRESSION, be able to write every book in existance into it,
and it would have a polymorphic writing personality, it would sorta be
the complete average of everything that went into it.
and its completely doable! the motion picture version is highly
challenging, but the word version? simples! code it for fun!
(the motion picture version is the super tv for the looney vision you
your supposed to show it nothing but repetitive classic cartoons and
then have fun watching it glue it all together on the way back out! :)
so we must tell the students about it!!!!
code it any way you want… doesnt have to be biologically exact or
anything i dont think.
but! it can tell you which writer WROTE IT! insane.
bit grouping. works on ANY DATASTREAM. especially simple to think about
with a character string stream, then one gets to catch phrases
and no… no rules of grammar magically appear in it.
but if it was motion video? it will tell you what goes with what in
still. very exciting thing, not sure how one would produce the robot
from it, but gosh get it working with video!!! gosh yes, jonny
cab from total recall (silly little finite state machine with symbol
senses) is a reality.
so, emulating human senses??? hmm, not sure if i would call it human
senses, but computer senses maybe.
well, youve got till 2018 to get it done, then ibm’s prediction was
Please log in or register to post a reply.