Post by Admin on Jun 1, 2015 18:03:55 GMT
There are a few things you can do to do this. First you need the Ai to
learn the words in the documents. To do this you would use the keyword
ReadFile to read the files to the input then if LM learning is ticked in
settings new words and their meaning will be harvested from the online
dictionary.
To set up a file to read to input:
mode = 0
file = filename of the file to read
keyword = ReadFile
field 7 = 1
note if field 7 is empty then nothing will happen.
The file will be read into the input line by line, each line must be
less than 240 characters long otherwise it will be rejected.
I'm not sure if a pdf file is text only you might need to copy and
paste the text into a text file first.
The RAS and Parameters routines both work with files that they use for
source data. The file name in field 11 of cell 819 holds the filenames
of files to use. This file holding filenames can be created manually or
automatically using the keyword IndexFiles.
The DT routine makes connections in the Brain between cells and trains
the neural network according to how you tell it to re-shape its previous
answer. if you say to the Ai "hello how are you" and it replies "fish
toe you" or some other gibberish then you can tell it what you would
prefer it to say using DT learning this is done by either typing your
preferred reply into the output box or by a pattern of words usually "no
say" triggering the keyword DTLearn. If typing into the output box press
ENTER then the previous input will point to the new words by altering
the in cell fields updating the neural network training file and then
using the training file to train the network. If the training sessions
all reach 100% then the exact words will be used, if this happens is
dependant on the max number of generations in settings and also how many
conditions need to be met. Each training session adds new data which
makes it harder and harder for the network to train to a full 100%. If
the Ai's outputs get set in there ways then the training data can be
deleted.
6p learning will also make associations but this function is heavy on
processing resources so I normally tick it in settings input a few
things let it do its thing then un tick it.
NNEB and NNEC also work on associations created by LM, DT, and 6p
I hope this helps and isn't to confusing if you need anything
clarifying or more help please let me know
learn the words in the documents. To do this you would use the keyword
ReadFile to read the files to the input then if LM learning is ticked in
settings new words and their meaning will be harvested from the online
dictionary.
To set up a file to read to input:
mode = 0
file = filename of the file to read
keyword = ReadFile
field 7 = 1
note if field 7 is empty then nothing will happen.
The file will be read into the input line by line, each line must be
less than 240 characters long otherwise it will be rejected.
I'm not sure if a pdf file is text only you might need to copy and
paste the text into a text file first.
The RAS and Parameters routines both work with files that they use for
source data. The file name in field 11 of cell 819 holds the filenames
of files to use. This file holding filenames can be created manually or
automatically using the keyword IndexFiles.
The DT routine makes connections in the Brain between cells and trains
the neural network according to how you tell it to re-shape its previous
answer. if you say to the Ai "hello how are you" and it replies "fish
toe you" or some other gibberish then you can tell it what you would
prefer it to say using DT learning this is done by either typing your
preferred reply into the output box or by a pattern of words usually "no
say" triggering the keyword DTLearn. If typing into the output box press
ENTER then the previous input will point to the new words by altering
the in cell fields updating the neural network training file and then
using the training file to train the network. If the training sessions
all reach 100% then the exact words will be used, if this happens is
dependant on the max number of generations in settings and also how many
conditions need to be met. Each training session adds new data which
makes it harder and harder for the network to train to a full 100%. If
the Ai's outputs get set in there ways then the training data can be
deleted.
6p learning will also make associations but this function is heavy on
processing resources so I normally tick it in settings input a few
things let it do its thing then un tick it.
NNEB and NNEC also work on associations created by LM, DT, and 6p
I hope this helps and isn't to confusing if you need anything
clarifying or more help please let me know