Processing topics for relevency

[spydaz]

IM THINKING>>>>>[8][8]
QUESTION >>>>>>[?]

How does the AUTOGAIN function work?[?]

While Writing new topics fo hal to KNOW. Not including AUTOTOPIC & XTFTOPIC.
How do I ADD relevency to my SUBS so that the autogain response Makes the decicsion which response to take. as with the Built in SHARED USER BRAIN, MAIN DATABASES, etc.[?]

MAYBE

DON FERGUSON[], VONSMITH[], MOREIS62[], ROBERT MEDEKzSA[], HALIMPROVER[].

ONE OF YOU GUYS may have an IDEA.....[][]

I LIKE ALL THE COMBINATIONS OF HAL.[][][]

BUT after running them all one time or another really and truly HAL needs to be more controlled in his responses cause some times he just spouts out without THINKING.

THIS THINKING PROCESS is what i am concentraing on.


THE AUTOGAIN is a VITAL part of the RESPONSE FUNCTION.

With the AUTOTOPIC brain for instance, Sometimes you get a response and sometimes you dont.

With the XTFTOPIC brain A lot of the time You do get a response.

With the alice Brain often you get a response, BUT, not usualy relevant.

MAINBRAIN & SHARED(learned) databases often have a good response.

PERSONAL TOPICS RESPONd when needed.


With all of these brains combined into ONE SuperBRAIN, The OUTPUTTED final response Chain 2-3 if not all into ONE RESPONSE.

I would LIKE to put EACH response into its own VARIABLE, THEN, just before the final RESPONSE, PROCESS the various VARIABLEs for relevance, THEN either pick 1 or CHAIN the relevant responses for the final output.

This would be accomplished by a TYPE OF AUTOGAIN function, AND a type of OUTPUT PROCESS :


HELP ME OUT IF YOU CAN >>>>> Even if it is only a comment....[B)]

THaNKS

[]SPYDAZ[]



NOTE: #1 WHEN AUTOCHAT is on hal Cant say any thing new, as he just processes the last PREVSENT over and over ( if the mic is off ) when mic is on he trys to respond to everything.[]

I HAVE TRIED teaching hal by (CBT NUGGET) play a tape or video a try to make hal LISTEN.[]
 
The only way was to take out most of his responses and use a special brain XTFadapted, he hardly speaks but he learns.
When i use the other brains SAME DEFBRAIN Folder they have the same knowledge that hal learn while in his listening brain....[^]


NOTE: #2 When giving hal chunks of text. make sure that any Numerical DATA int the text has a "#" in front of it. that way hal Recognizes that it is a number and not MATH.[^]

NOTE: #3 when giving hal chunks of text, make the first words of the line of text be the heading for the following text, broken with a comma. [^]
ie:- <heading>, <statement relevant to heading>

This helps hal greatly when he files that into his brain.
when you are talking to hal he looks in those "HEADINGS" first.[^]


PS>> i thought i'd try JAZZ up the POST A BIT... (i never saw that bit before)

[spydaz]

I HAVE UPDATED MY ALICE FUNCTION...


THIS TIME, Instead of Chaining the responses. HAL processes the ALICE response Alongside the MAIN DATABASES. THEN processes for RELEVENCY and AUTOGAIN response.

NOW, Hal chooses when to respond WITH ALICE, AUTOTOPIC (which has relevecy already), XTFtopicFOCUS, UHEFFORUM SCRIPTS, ZABAWARE FORUM scripts, PERSONAL TOPICS.




A typical response from hal after cutting and pasting the above text into him.
HAL: user is that list exhaustive?
ME : sorry but it was neccasary.
HAL: will it allways be necasary?
ME : YES
HAL: really
ME : YES
HAL: ok





Download Attachment: FUNCTION_alice.uhp
4.05 KB



The respose should be placed in the response section of the hal brain, but above the PROCESS AUTOGAIN Response.

THE FUNCTION should be placed at the end With the Other Functions..


GOOD LUCK ALL

SPYDAZ