[X]Assistant

[shahmatist2]

At last, I found a program for the obshalku, which did not even hit the counterpart on this site. It is called Alica. In addition to communicating in text mode, he still knows how to manage the PDA system. But one minus - she only speaks English. Is it possible to remake it in Russian?
I also constantly work on my iSiN project. Below I post the source code on basic4ppc.

Jacco_Bikker_AliCE.cab(415.67 KB)
iSiN.zip(1.54 MB)


Here, only a general theoretical discussion!
Topics for discussing programs:

iSiN
Alice
STEB

Post has been editedshahmatist2 - 16.06.12, 09:46

Reason for editing: added links and warning

[shahmatist2]

Improved AI loading algorithm
Here is:
version
6.50
0
Form1
0
Sub designer
addform (Form1, "Form1", "", 211,211,211) @
End sub
@ EndOfDesignText @
'#######THE TASK#######
'# GLOBAL VARIABLES #
Sub globals
Dim MatricSinaps (0,0 0,0), BlokSinaps (0) as String
End sub
'____________1______________
'#####INITIALIZATION#####
'###### INTELLIGENCE #######
Public Sub START
obrabotca_documentov
End sub
'____________2______________
'### FILE TREATMENT ###
Sub obrabotca_documentov
'# (Assigning paths to variables) #
DIRintellekt = AppPath & "\ INTELLEKT \"
EXT = ". Wrd"
'# (Activation of dynamic arrays)
AddArrayList ("Alfavit")
AddArrayList ("Themes")
AddArrayList ("iSiN_info")
AddArrayList ("Chelovek_info")
AddArrayList ("Set_memory")
AddArrayList ("Log")
AddArrayList ("Input_patterns")
AddArrayList ("Output_otvet")
AddArrayList ("Temp")
AddArrayList ("Flag")
'# (Assigning memory to a cell) #
'# ((((((((((((((of the alphabet)))))))))))))) #
FileSearch (Alfavit, DIRintellekt & "WORD", "*" & EXT)
Lenth = Alfavit.Count-1
For i = 0 to Lenth
Alfavit.Item (i) = StrReplace (StrReplace (Alfavit.Item (i), DIRIntellekt & "WORD \", ""), EXT, "")
Next
'# (Creating dynamic arrays) #
'# (((((from the memory of the cell of the alphabet)))))) #
For i = 0 to Lenth
AddArrayList ("word" & i)
AddArrayList ("sinonim" & i)
Next
'# (Algorithm handling synonyms in) #
'# ((((((((((((((((((((((RAM))))))))))
For i = 0 to Lenth
TEMPname = Alfavit.Item (i)
If FileExist (DIRintellekt & "WORD \" & TEMPname & EXT) Then
If FileExist (DIRintellekt & "sinWORD \" & TEMPname & EXT) Then
FileOpen (c1, DIRintellekt & "WORD \" & TEMPname & EXT, cRead)
FileOpen (c2, DIRintellekt & "sinWORD \" & TEMPname & EXT, cRead)
For j = 0 to 999999
TEMPslovo = FileRead (c1)
TEMPsinonim = FileRead (c2)
If (TEMPslovo = EOF) OR (TEMPsinonim = 0) Then
EXIT
End if
Control ("word" & i) .Add (TEMPslovo)
Control ("sinonim" & i) .Add (TEMPsinonim)
Next
FileClose (c1)
FileClose (c2)
Else
MsgBox ("Install base of synonyms")
End if
Else
MsgBox ("Install the database of words")
End if
Next
'# (search for conversations) #
FileSearch (Themes, DIRintellekt & "THEMES", "*" & EXT)
'# (naming) #
'# (to those dynamic arrays) #
Lenth = Themes.Count-1
For i = 0 to Lenth
AddArrayList ("themes" & StrReplace (StrReplace (Themes.Item (i), DIRintellekt & "THEMES \", ""), EXT, ""))
Next
'# (assignment of data to themes) #
'# (corresponding to the given arrays) #
For i = 0 to Lenth
TEMPname = Themes.Item (i)
FileOpen (c1, TEMPname, cRead)
Themes.Item (i) = StrReplace (StrReplace (Themes.Item (i), DIRintellekt & "THEMES \", ""), EXT, "")
For j = 0 to 999999
TEMPslovo = FileRead (c1)
If TEMPslovo = EOF Then
EXIT
End if
Control ("themes" & Themes.Item (i)). Add (TEMPslovo)
Next
FileClose (c1)
Next
'# (Reading info isina or creating info)
If FileExist (DIRintellekt & "iSiN" & EXT) Then
FileOpen (c1, DIRintellekt & "iSiN" & EXT, cRead)
For i = 0 to 999
TEMPslovo = FileRead (c1)
If TEMPslovo = EOF Then
EXIT
End if
iSiN_info.Add (TEMPslovo)
Next
FileClose (c1)
Else
TEMPinfo = ""
MatricSinaps () = Array (("Name", "Isinka", "Eve", "Alice", "Light", "Julia", "Tatiana", "Oksana", "Marina", "Elizaveta", "Lyudmila" , "Hope", "Margaret", "Sasha", "Love", "Ekaterina"), ("Age", "15", "16", "17", "18", "19", "20" , "21", "22", "23", "24", "25", "26", "27", "28", "29"), ("Growth", "161", "162" , "163", "165", "166", "167", "168", "169", "170", "171", "172", "173", "174", "175", " 176 "), (" Breast size "," 2 "," 3 "," 4 "," 5 "," 2 "," 3 "," 4 "," 5 "," 2 "," 3 ", "4", "5", "2", "3", "4"), (Likes when it is called, "" fish "," kitten "," sun "," angel "," Isinka ", "Murka", "bunny", "cutie", "darling", "clever", "baby", "clever", "beloved", "little", "girl"), ("Hair color", "blonde" , "blonde", "brunette", "brunette ka "," shotenka "," shotenka "," redhead "," redhead "," brunette "," brunette "," blonde "," blonde "," shotenka "," redhead "," redhead "), (" Eye color "," brown "," green "," blue "," gray "," blue "," nacreous "," dark green "," dark blue "," dark gray "," light green "," pearl-brown "," dark brown "," light blue "," gray-green "," brown "), (" Direction of music "," pop "," rock "," pops "," rap, classic, hip-hop, rnb, techno, electro, instrumental, metal, metal, rock, pop, punk rock ), ("Cinema Direction", "Melodrama", "Thriller", "Action", "Fantasy", "Crime", "Comedy", "f Antastik thriller "," musical "," historical "," historical thriller "," comedy melodrama "," horror "," psychological thriller "," military "," Soviet "), (" Favorite color "," blue ", "black", "red", "green", "blue", "white", "yellow", "gray", "lilac", "nacreous", "orange", "brown", "lime", "golden" "," silver "), (" Favorite topic of conversation "," sex "," love "," meaning of life "," happiness "," science "," life "," mystic "," zone "," music " , "cinema", "study", "entertainment", "fate", "death", "jokes"))
For i = 0 to ArrayLen (MatricSinaps ()) - 1
TEMPslovo = MatricSinaps (i, Rnd (1,15))
iSiN_info.Add (TEMPslovo)
TEMPinfo = TEMPinfo & MatricSinaps (i, 0) & ":" & TEMPslovo & crlf
Next
TEMPinfo = SubString (TEMPinfo, 0, StrLength (TEMPinfo) -2)
FileOpen (c1, DIRintellekt & "iSiN" & EXT, cWrite)
FileWrite (c1, TEMPinfo)
FileClose (c1)
End if
'# (Reading the information of the head)
If FileExist (DIRintellekt & "Chelovek" & EXT) Then
FileOpen (c1, DIRintellekt & "Chelovek" & EXT, cRead)
For i = 0 to 999
TEMPslovo = FileRead (c1)
If TEMPslovo = EOF Then
EXIT
End if
Chelovek_info.Add (TEMPslovo)
Next
FileClose (c1)
Else
MatricSinaps () = Array (("Name", "Age", "Growth", "Breast size", "Likes when it is called", "Hair Color", "Eye Color", "Direction of Music", "Direction of Cinema "," Favorite color "," Favorite topic of conversation "))
TEMPinfo = MatricSinaps (0,0) & ":" & "Man"
Chelovek_info.Add (TEMPinfo)
TEMPinfo = TEMPinfo & crlf
For i = 1 to 10
TEMPslovo = MatricSinaps (0, i) & ":" & "Don't Know"
Chelovek_info.Add (TEMPslovo)
TEMPinfo = TEMPinfo & TEMPslovo & crlf
Next
TEMPinfo = SubString (TEMPinfo, 0, StrLength (TEMPinfo) -2)
FileOpen (c1, DIRintellekt & "Chelovek" & EXT, cWrite)
FileWrite (c1, TEMPinfo)
FileClose (c1)
End if
'# (Reading conversation memory)
If FileExist (DIRintellekt & "Set_memory" & EXT) Then
FileOpen (c1, DIRintellekt & "Set_memory" & EXT, cRead)
For i = 0 to 999
TEMPslovo = FileRead (c1)
If TEMPslovo = EOF Then
EXIT
End if
Set_memory.Add (TEMPslovo)
Next
FileClose (c1)
Else
FileOpen (c1, DIRintellekt & "Set_memory" & EXT, cWrite)
Set_memory.Add ("Offended: no")
Set_memory.Add ("I was delighted: no")
Set_memory.Add ("Distressed: no")
Set_memory.Add ("I learned: no")
FileWrite (c1, "Offended: no" & crlf & "Delighted: no" & crlf & "Distressed: no" & crlf & "Learned: no")
FileClose (c1)
End if
End sub
'# MatricSinaps Logic Matrix
'# BlokSinaps partitioning array
'# Alfavit-alphabet (files a.wrd, b.wrd, ... i.wrd contain the corresponding bases of words)
'# word & n-words. They are in my databases somewhere 98000
'# sinonim & n-synonyms. There are as many of them as there are words.
'# Themes talk topics (titles)
'# Themes & n-text information of topics (the algorithm for working with this information is still under development)
'# iSiN_info info isina
'# Chelovek_info-info man
'# Set_memory memory of the last conversation
'# Log Log Entry
'# Input_patterns-input patterns
'# Output_otvet-ready answer
'# Temp-temporal dynamic array
'# Flag variable tags

[shahmatist2]

The person of Isin is created when the algorithm is first run, and then does not change. The parameters of the person Isin are not unequivocally chosen, i.e. she can become either Lena with some preferences or Eve with others. Unambiguity is only that it will be a girl :)

[shahmatist2]

Chelovek_info array will be filled during the conversation, and saved to a file when you close isin. I have one idea, so that Yishin would apply to “You” until it fills this array or the user doesn’t say that it’s time to switch to “you”, and to fill it, he asked questions if they fit the topic.

[shahmatist2]

Added extended database from talkerus desktop program

Attached files

iSiN.zip(1.54 MB)

[Alt]

You do an interesting thing, but it seems that the implementation tool chose not a very good one. Still, Basic4ppc is pretty limited development environment. Sooner or later, creating a serious project, the developer rests on a huge pile of stupid restrictions, which are pretty hard to overcome.

[shahmatist2]

For word processing, bye this one will come down. Then you have to learn c ++.

Posted on 04/27/2012, 1:34 PM:

The main feature of basic4PPC is in the visualizer, but I do not work with it, because the interface in this program is almost irrelevant.

[shahmatist2]

And in general, while on what to write is not important. It is necessary to develop new algorithms, even if they think. I've encountered such a problem:
How to create an algorithm for deciphering the meaning of text without the key patterns of these meanings. There is an idea of ​​classifying the text on sentences, then on words and a further search for the number of matches for each word. The winner is the word that most often comes across in the text, i.e. it is key in the sense of the latter. Then look for the word that will be in second place by the number of matches, etc. These words drive into an array of priorities. What next to do, I have not yet invented.

[Taifunfly]

I checked again ... While it’s not interesting to talk to him / her ... "Man, let's talk about the weather?" I start writing about the weather, he offers me to talk about nature ... etc.)) The only thing I liked was when I write a smiley ":-)" or any other, he asks me ":-)" what is it? I think in there is a reason ... let it be in the form of a small child who asks "what is it?" \ "something?", "and why so?", and when you answer him, he starts learning how to recognize you and how peace ... You can also make an opportunity to fill in your parameters initially (before communication), your profile, so that this "Isina" already knew you and your preferences, hobbies ..
And then she will ask you: what is music? What is "disco", well, etc., and fill your answers into your "Artificial Brain Base" ... So gradually, if your (established at your km / kpk ) "Artificial Intelligence" will train the forum users and spread the filled base with their own knowledge .. a normal AI will be assembled with the world, who can speak on different topics ... You will talk about music with him, he will answer: "Artem from Volgograd said that music needs to be felt and not everyone understands it ... "Well, this is just my IMHO ...

[shahmatist2]

Taifunfly @ 29.5.2012, 22:31

I checked again ... While it’s not interesting to talk to him / her ... "Man, let's talk about the weather?" I'm starting to write about the weather, he offers me to talk about nature .. etc.))

With iSiN 1.2 it is more interesting to communicate and it defines the topics of conversation. About learning and questionnaires, our ideas converge. It’s just that I’m still stuck in a very difficult way - these are self-learning algorithms (even without a companion, for example: reading a book or an incyclopedia :)).

[shahmatist2]

There is the idea of ​​a new project Logic Sistem Exploer (Logical System Explorer). Desired features:
1. Multiple windows, like the conductor of a big brother. (There are ideas).
2. Multitasking. For example, in one window you can simultaneously insert and delete files. (There are ideas).
2. The ability to drag and drop files from one window to another (copy). (No ideas yet :().
3. Group opening of files. (There are ideas).
4. Disk visualization (C: -Memory device, D: -Memory card). (Got ideas).
4. Virtual assistant. (Yes).
5. Scanning and notification of changes in the file system when starting LSE. (There are ideas).
6. All features Resco Exploer. (There are ideas, except plug-ins).

[123456ololo]

you can try to create a self-learning neural network, though the problem will be to implement it in basic

[shahmatist2]

On this basic even the HTM (Self-Learning Temporal Memory Grouper) can be made. It is even more abrupt than neural networks :)

[123456ololo]

shahmatist2,
Well, if so, then you can try to make NA

[shahmatist2]

For HTM, only operations for working with large arrays, matrices, and stacks are needed, as well as support for creating subroutine functions. About a large array: I tried to create a dynamic array on basic4ppc
AddArrayList ("BigSize") for which BigSize.Count-1 = 186000 (one hundred and eighty six thousand word-filled elements!), Which filled in its elements from the lines of a text file. Loaded really about 7 minutes. An HTM of more than 186,000 cells is no longer a “frail” brain.

Post has been editedshahmatist2 - 11.06.12, 08:14

[hitmen047]

As a programmer with good basic knowledge, I will contribute to the conversation.
In fact, advanced AI relies on a kind of database. Therefore, the entire structure should be in the form of a large dynamic database. I am sure that I haven’t told you anything new now, however, I can confidently say that quite a lot of well-known C # language works best for working with databases. Which in turn interacts with the .Net Framework and has an excellent supply of auxiliary libraries, which facilitates the perception of the syntax as a whole. Given that Win Mobile supports the Compact Framework (such as before version 3.5), C # is the best way to accomplish this task. Also, besides everything, is the database better than the file system? The fact that due to regular expressions you can easily refer to a particular data area. From this it can follow that the perception of the SQL server is also required.
Thus, summing up everything I said above, leads to the fact that, after all, but the solution will be SQL Server and C #. It remains only to find a programmer who agrees to accomplish the task. And yes. Once the AI ​​will communicate with the person, it would not be bad to teach him. That bij, to enter the new data in a DB. From here you can dig for a long time in the direction of a single data server on which information could be stored. But this is a completely different and more difficult task ...

[shahmatist2]

For me the choice of the best way to implement the ideas of stock in second place. The main thing is to have the POSSIBILITY of their implementation, so that the algorithms themselves are designed (if they will be called that after this). I have one idea for creating a “special” network in which the database plays a secondary role ... I cannot even express the algorithms of this network yet. She is only in theory with a couple of formulas.

[AlkexeyP]

And yet, why the PDA?

[hitmen047]

AlkexeyP @ 06/12/2012, 12:32

And yet, why the PDA?

Probably because the forum section on programming for PDAs. ;)

shahmatist2 @ 06/12/2012, 04:26

For me, the choice of the best way to implement the ideas of stock in second place. The main thing is to have the POSSIBILITY of their implementation, so that the algorithms themselves are designed (if they will be called that after this). I have one idea for creating a “special” network in which the database plays a secondary role ... I cannot even express the algorithms of this network yet. She is only in theory with a couple of formulas.

Oh ... Moderators will punish me for this. But as an aid to programming I can offer an excellent forum,Cyberforum. There are programmers \ system administrators \ PC lovers, etc.

[shahmatist2]

AlkexeyP @ 12.6.2012, 16:32

And yet, why the PDA?

The PC already has a bunch of similar Russian-language prog. On the PDA for WM - not one. That's why I took up this business. Especially if you get a good AI simulator, then the most functional place for the implementation of the PDA. Anywhere, in the subway, car and just on the road, you can talk and ask advice from the AI.

[Taifunfly]

Greetings to the author and the developer of the AI ​​on the PDA. I would like to know how things are progressing towards improving this great idea?