dupa

Author Topic: SUCESS Stage 1  (Read 4561 times)

spydaz

  • Hero Member
  • *****
  • Posts: 670
    • View Profile
    • http://www.spydazweb.co.uk/
SUCESS Stage 1
« on: April 04, 2006, 11:43:54 am »
I have now written a program to extract the wordnet subsets,

MERONYMS
HYPERNYMS
SYNONYMS
SISTERS
Definitions.

Into a DB.
well 5 mini 2 column dbs


Private Sub CMD_Read_meronyms_Click()
Dim wn As New WordNet
fileNAME = App.Path & "\" & txt_Filename.Text

LineNum = GetFileLength(fileNAME)
txtfileOUT.Text = LineNum


'Search file Line by line
'   Lookup in word net
'   get list of meronyms
'   * TEMP : display list in output win
' remove ( )
' How many meronyms
' Get meronym 1 by 1
' update DB with NOUN (FROM FILE) & MERONYM FROM LIST
' return


For CurrentLine = 1 To LineNum

'Search file Line
Noun = C_ReadLINE(fileNAME, CurrentLine)

'Lookup in word net
'This points WN to your word "horse".
wn.LookupWord (Noun)

'   get list of meronyms
List_Of_Meronyms = ""
List_Of_Meronyms = wn.GetMeronyms(1) & "  " & wn.GetDefinition(wn.GuessPartOfSpeech, 1, "S")

'   * TEMP : display list in output win
'txtfileOUT.Text = List_Of_Meronyms


' Get meronym 1 by 1
Do While List_Of_Meronyms <> ""
Meronym = C_GetFirstWord(List_Of_Meronyms, contents, ",")
' contents contains the rest of the list
If contents <> "" Then
    List_Of_Meronyms = contents
    Call DBMeronymUPDATE(Noun, Meronym)
   
 Else
List_Of_Meronyms = ""
Meronym = ""
End If
Loop


Next



End Sub


currently a ms access db, this can now be cut and pasted across to the HAL6 DB.....


first i have a list of NOUNS in a text file, then the SUB searches the textfile line by line. looking up every word and extracting is *NYM and updating the DB TABLE accordingly, NOUN/MERONYM  NOUN/HYPERNYM NOUN/SISTER etc.

a step towards Learning from a text file....

Next i will be concentrating on Learning new phrases from text files, as in writers when characters are Speaking;
"hello" said harry
" whats up said KIM"

there speech is contained within Speech MARKS. these Quotes could be stored as Potential QUESTION AND ANSWER phrases, AS WITH F.A.Q.

this will enable our hal's to learn from conversation logs from Other AI conversations, even POP STAR/ WWE interveiw transcripts or TV show "Lines" could be learned, as QUESTION AND ANSWER just for those guys with all questions ever asked to DATA on Star treK and Answers......


A Possiblity... A potential Charcter builder ( not Knowledge )


spydaz

  • Hero Member
  • *****
  • Posts: 670
    • View Profile
    • http://www.spydazweb.co.uk/
SUCESS Stage 1
« Reply #1 on: April 08, 2006, 05:54:51 am »
1. Problem.


I have come upon a problem with the GETSISTERS Even in hal / VBSCRIPT

When calling for getsisters the Whole program / script crashes ..

Ususally you would get a RUNTIME Error code but you just get a SEND ERROR REPORT... Breeds no explanation...

I wonder If anyone has had a problem wih this routine... Help If Possible....

i have tried multiple NOUNS and they all crash.

Is the sisters routine only for nouns or verbs or adjectives ?

Or is it a Poke / ERROR in the Routine Itself I am Using the HALWN.Dll to extract the data....

I shall post the Exported textr for the Successfull wordnet extractions TODAY...

These can be Imported to the HALBRAIN.DB using the braineditor....

Hopefully we can end up REMOVING the REAL wordnet from HAL Inplace or the NEW Imported DATA ... Giving US the Ability to build up and EDIT our OWN dictionary inside HALS brain...... Simularly i also have VARIOUS EURO Wordnets  ...
By using the ENGLISH version as a base I have the NOUNS / VERBS / ADVERB etc in FRENCH / SPANISH / ITALIAN ..

This will enable hal to become Multi Lingual ...

As recieving the english phrase and Processing LANGUAGE substiutions for EACH word in the SENTENCE ...#




REAL KNOWLEDGE / REAL SENSE /REAL COGNITION is my Final GOAL (Simulated EMOTIONS)

spydaz

  • Hero Member
  • *****
  • Posts: 670
    • View Profile
    • http://www.spydazweb.co.uk/
SUCESS Stage 1
« Reply #2 on: April 09, 2006, 04:43:44 am »
I cant upload ????

Medeksza

  • Administrator
  • Hero Member
  • *****
  • Posts: 1472
    • View Profile
    • http://www.zabaware.com
SUCESS Stage 1
« Reply #3 on: April 09, 2006, 09:38:58 am »
I haven't downloaded or tried any of the following myself, but I came accross the following projects and may consider using one of them in future versions of Hal in order to upgrade to WordNet 2.1 (Hal uses WordNet 1.7)

http://www.androidtech.com/html/wordnet-mysql-20.php
http://sourceforge.net/projects/wnsqlbuilder/
http://wordnet.princeton.edu/links.shtml
Robert Medeksza

spydaz

  • Hero Member
  • *****
  • Posts: 670
    • View Profile
    • http://www.spydazweb.co.uk/
SUCESS Stage 1
« Reply #4 on: April 29, 2006, 05:19:14 am »
funnyly enough MR M. i tried that whole procedure/ prolog wordnt.

it gave me the idea of how to extract the db.

But as you can see with the above programming, ( You have already completed most of the programming work already ) in your HAL DICTIONARY. it could be adapted to read our current wordnet 1.7 / ? 2.0 >> into the haldb.

its just the form which the TABLE will be in ( Thats the question) as i have made various Large data tables (all usefull) retrieving the data is a another question  as some of the QUERYS do take a little while to execute (Not on my dual core) but on some machines (you get the old THIS SCRIPT IS WAITING TO EXECUTE) wow, i thought it was lost from hal5....


onthecuttingedge2005

  • Guest
SUCESS Stage 1
« Reply #5 on: April 29, 2006, 02:23:00 pm »
quote:
Originally posted by spydaz

funnyly enough MR M. i tried that whole procedure/ prolog wordnt.

it gave me the idea of how to extract the db.

But as you can see with the above programming, ( You have already completed most of the programming work already ) in your HAL DICTIONARY. it could be adapted to read our current wordnet 1.7 / ? 2.0 >> into the haldb.

its just the form which the TABLE will be in ( Thats the question) as i have made various Large data tables (all usefull) retrieving the data is a another question  as some of the QUERYS do take a little while to execute (Not on my dual core) but on some machines (you get the old THIS SCRIPT IS WAITING TO EXECUTE) wow, i thought it was lost from hal5....





You could always setup a Stop Watch Function at the very top of the Brain and a Stop Watch Function at the Bottem of the Brain so that if there is a difference in Seconds between the top Stop Watch and the Bottem Stop Watch then you could use the Exit Function to execute so the Wscript Time out will not occure Since the reading of code always starts from top to bottem literally.

There are many ways of setting up different timers to monitor code speed so that if a Script is taking a long time then it could close out the Function to prevent the Time out from occuring especially in Arrays.

Here are some custom Time Functions:

Code: [Select]

'Display the Seconds of a Clock.
Function MyWatch()
   If MyWatch = "" Then MyWatch = HalBrain.SearchPattern(FormatDateTime(Time(), vbGeneralDate), "*:*:* *", 3)
End Function

'Or

'Display Last Time since Midnight.
Function MyStopWatch()
   MyStopWatch = Timer
End Function


The Timer Possiblities are endless and the Exit Function is very handy indeed.

Jerry[8D]