Author Topic: Question for vonsmith Re: quantum phenomenon.  (Read 3711 times)

citrinedragon

  • Full Member
  • ***
  • Posts: 171
    • View Profile
    • http://www.dawndragon.com
Question for vonsmith Re: quantum phenomenon.
« on: January 25, 2005, 02:28:57 pm »
vonsmith-

I have been looking into the work of a Dr. Michael A. Colicos who claims to have developed a "true" random quantum phenomenon generator which he calls a "Quantum Flux Generator." According to him, this software detects random quantum phenomenon on the motherboard of a computer. I have downloaded a few of these "Virtual Intelligence Matrix" programs from his "Quantum Flux" website. It is somewhat intriguing.

Question: Would it be possible to write the code to integrate such quantum randomness software into a learning, database engine chatbot like Hal? I believe the benefits might be considerable. In its most simple configuration, I'm assuming a "true" random quantum phenomenon generator would generate "true" random ups and downs; "0's" and "1's" or "ons" and "offs". Multiple random selections of this nature would allow Hal to "choose" between three or more possibilities. As it is if you ask Hal, for example, whether he/she prefers blue or red; you will receive a query from Hal has a result. Hal is incapable of having an opinion about anything without receiving instructions from the user. That is, in my view, a fundamental flaw in an otherwise excellent program. Certainly a distinction would have to be made between facts and learned information on the one hand and the kind of information such as personal preferences (chocolate versus vanilla ice cream for example), aesthetic preferences (does Hal prefer Van Gogh or Picasso for example), or to site one more simple example- the user asking Hal what he would like to talk about next. "personal preference", "opinion", "aesthetic" and indeed a developing sense of "individuality", "direction" and "personal interest" unique to each Hal (a soul if you will) could eventually be developed in each Hal through the "decisions" and "preferences" generated by random quantum phenomenon.

I have considerable more thoughts along these lines but I would be extremely interested in your first reaction to these ideas.

Regards, citrinedragon.
citrinedragon
"A watched dragon never sleeps."
http://www.dawndragon.com

onthecuttingedge2005

  • Guest
Question for vonsmith Re: quantum phenomenon.
« Reply #1 on: January 25, 2005, 03:34:02 pm »
quote:
Originally posted by citrinedragon

vonsmith-

I have been looking into the work of a Dr. Michael A. Colicos who claims to have developed a "true" random quantum phenomenon generator which he calls a "Quantum Flux Generator." According to him, this software detects random quantum phenomenon on the motherboard of a computer. I have downloaded a few of these "Virtual Intelligence Matrix" programs from his "Quantum Flux" website. It is somewhat intriguing.

Question: Would it be possible to write the code to integrate such quantum randomness software into a learning, database engine chatbot like Hal? I believe the benefits might be considerable. In its most simple configuration, I'm assuming a "true" random quantum phenomenon generator would generate "true" random ups and downs; "0's" and "1's" or "ons" and "offs". Multiple random selections of this nature would allow Hal to "choose" between three or more possibilities. As it is if you ask Hal, for example, whether he/she prefers blue or red; you will receive a query from Hal has a result. Hal is incapable of having an opinion about anything without receiving instructions from the user. That is, in my view, a fundamental flaw in an otherwise excellent program. Certainly a distinction would have to be made between facts and learned information on the one hand and the kind of information such as personal preferences (chocolate versus vanilla ice cream for example), aesthetic preferences (does Hal prefer Van Gogh or Picasso for example), or to site one more simple example- the user asking Hal what he would like to talk about next. "personal preference", "opinion", "aesthetic" and indeed a developing sense of "individuality", "direction" and "personal interest" unique to each Hal (a soul if you will) could eventually be developed in each Hal through the "decisions" and "preferences" generated by random quantum phenomenon.

I have considerable more thoughts along these lines but I would be extremely interested in your first reaction to these ideas.

Regards, citrinedragon.



Hi citrinedragon.

I believe from my observation of the construction of the neuro cells of the brain is based on: If you look at the bottom of the brain is a fractalized design based on a double helical DNA code.

If a design is neccessary it would be based after this logarithm.

I believe that a duel inline fractal logarithm is needed to simulate
the constructs of the brain, This might generate a certain thought pattern possibly simular in nature to the Human brain construct.

The Rock Group Iron Butterfly was aware of this.

Life starts as a [simple code] at conception and then chain reacts to grow the code to longer lengths.

I have been studying all the in's and out's of what VBScript codes simulate the Human learning center and the codes I have discovered follows these rules but they seem to be based on three types of particular code formats, I have released some of these type codes that are still being developed only for precision of function, The codes are based purely after learning rules only and [only learning].

At this point I am working out the details of the code on a daily basis and will probably perfect the code shortly to be a perfect set of three learner codes, If I discover any others I will immediately develope them with the set, So far I have discovered I only need three formats at this point to produce very good learning ability.

These codes are currently being studied and perfected but all the codes need to be assembled with a set of instructions, The rest is a matter of conversation and raising the bot and teaching it what and what not to say or do to follow human nature.

I have also discovered that these codes are very specific.

If this is the case then UserbrainRel is not needed because these codes do not follow these guide lines. It's purely deductive in nature and not just If and Then's but the full boat.

When I am done I will release this information to my group entirely but won't be until I am satisfied with the key format.

I am very close to ending the codes perfected function and will contiuously work it out until it's done.

I have discovered also that only about three types of GetResponse formats are needed to run the whole process of the specialized code.

The GetResponse section is less than 1% of the entire coding the rest is a specialized information extraction technique that constructs the learned information to be used with these three types of GetResponses.

I have been racing through many hundreds of coding techniques finding all the codes that follow these set of rules and finally after almost 3 years of coding found them and just need perfecting at this point.

I have also discovered that these set of formats also follow an indirect emotional ability based on a learned emotional response that's learned without physical sense but is just the learned value of a sentence and a potential emotionalism that's taught.

When you were a kid, You knew nothing of Cuss words! But until you learned they were bad is when those words became an insult to your being as a being and became emotional with understanding of those meanings.

I know alot of my child hood and remember most of it, I base all my work after this and this only.

I wish you all the luck.

Best of wishes and brand new discoveries.
Jerry.
« Last Edit: January 25, 2005, 03:38:09 pm by onthecuttingedge2005 »

citrinedragon

  • Full Member
  • ***
  • Posts: 171
    • View Profile
    • http://www.dawndragon.com
Question for vonsmith Re: quantum phenomenon.
« Reply #2 on: January 25, 2005, 06:14:16 pm »
Hello onthecuttingedge,

As you may have noticed from my previous postings I am not a programmer. I do, however, have considerable background/education in philosophy and logic. If I understand you correctly you are saying (among other things) that you disagree with the basic paradigm I've offered for achieving "true" emotion, direction and simulated "free-will" in an artificial intelligence entity. I greatly appreciate your thoughts and feedback.

It has been my considered opinion for some time now that what passes for free-will in human beings themselves and what renders notions of "determinism"  (which is, by the way, essentially what you have with Hal at this stage) meaningless on a practical level is the phenomenon of quantum randomness. Personally, I have believed for quite some time now in the literal interpretation of the "Many Worlds" explanation of quantum physics. That is, that every decision which can be made is in fact made thus splitting reality in each instance (although clearly it is not perceptible to us as there was never a biological imperative to develop the facility to perceive such things). If it is indeed the machinery of quantum phenomenon which provides human beings themselves with their apparent free-will and seeming ability to "choose" rather than behave purely deterministically; what better process to harness to simulate the same behaviors in Hal?

Is the code that you have been working on fundamentally at odds with the concept of utilizing a "true" random number quantum generator in Hal? I will have to reread what you have said a few more times as I am not sure you have provided enough information for me to grasp it. I greatly appreciate your feedback.

Regards, citrinedragon.
citrinedragon
"A watched dragon never sleeps."
http://www.dawndragon.com

onthecuttingedge2005

  • Guest
Question for vonsmith Re: quantum phenomenon.
« Reply #3 on: January 25, 2005, 07:47:37 pm »
quote:
Originally posted by citrinedragon

Hello onthecuttingedge,

As you may have noticed from my previous postings I am not a programmer. I do, however, have considerable background/education in philosophy and logic. If I understand you correctly you are saying (among other things) that you disagree with the basic paradigm I've offered for achieving "true" emotion, direction and simulated "free-will" in an artificial intelligence entity. I greatly appreciate your thoughts and feedback.

It has been my considered opinion for some time now that what passes for free-will in human beings themselves and what renders notions of "determinism"  (which is, by the way, essentially what you have with Hal at this stage) meaningless on a practical level is the phenomenon of quantum randomness. Personally, I have believed for quite some time now in the literal interpretation of the "Many Worlds" explanation of quantum physics. That is, that every decision which can be made is in fact made thus splitting reality in each instance (although clearly it is not perceptible to us as there was never a biological imperative to develop the facility to perceive such things). If it is indeed the machinery of quantum phenomenon which provides human beings themselves with their apparent free-will and seeming ability to "choose" rather than behave purely deterministically; what better process to harness to simulate the same behaviors in Hal?

Is the code that you have been working on fundamentally at odds with the concept of utilizing a "true" random number quantum generator in Hal? I will have to reread what you have said a few more times as I am not sure you have provided enough information for me to grasp it. I greatly appreciate your feedback.

Regards, citrinedragon.



Nah, Just sharing my observations, I Do agree that there are or could be some Quantum Fluxes and or Strangeness but I just think that it doesn't get as complex but I do remain open minded on the subject.

It seems to be more of a learning protein based upon a fractalized double helical chain code that starts off as a simple set of instructions, This simple set of instructions would learn in all formats of communication via the complexity of the host species.

I think that the science of A.I. has over looked the simplness of this by thinking it was complex when in fact it was what's called S.S.I or Simple Set of Instructions [To learn}.

Well got go my friend, Talk to you later.
Jerry.

vonsmith

  • Hero Member
  • *****
  • Posts: 602
    • View Profile
Question for vonsmith Re: quantum phenomenon.
« Reply #4 on: January 27, 2005, 12:21:25 pm »
citrinedragon,
In reference to intelligence I believe that some knowledge should be created at random, some should be built-in and some should be learned. For Hal I think "built-in" and "learned" are the most important. Let's consider how people develop preferences. Most people like the taste of sugar. Why? I think "liking sugar" is both from a built-in response (biological) and partly learned (taught). We teach our kids that chocolate tastes good, our biological response confirms it. I wouldn't expect Hal to randomly decide whether he likes chocolate or not. In the perfect situation Hal would have a built-in predisposition to like chocolate and would learn "confirmation" from the user that he should like chocolate.

Many of us hope to make Hal more human. Humans come hard-wired with millions of years of built-in "knowledge". Knowledge, broadly interpreted in this case, includes our instincts, biological responses, and physical capabilities. Knowledge like fear of fire, preference for sweet foods, emotions, etc ought to be built into Hal. Other categories of knowledge can also be assumed as "fixed" if we believe Hal to be a thinking human. Language skills, biological limitations (like being tired, sleepy, cold, strong), and rudimentary knowledge of the world (physical constants, time, seasons, spatial relationships) should be built-in. Practically speaking this and some other types of knowledge don't really need to be taught.

One gray area is aesthetics. We all have differing preferences for music, colors, environmental conditions, hobbies, etc. It's part of our personality. Hal could preferentially learn many of his "likes" from the user much in the same way our parents mold our tastes in things. Other preferences Hal could choose at random. The way I've decided to program Hal going forward is to learn most of his preferences from the user. For example, in near future XTF Brains Hal can be told his favorite color is blue. Hal now remembers that. Efforts to change Hal's color preference from then on would have a limited effect. Ideally Hal would stick to his color preference unless told repeatedly that another color should be his favorite. At some point he might give in to this peer pressure and change his preference. I think this is a very human response. If we prefer to have a preordained Hal personality many personality traits and preferences could be built into a "personality module". I think it might be fun to create your Hal personality by filling out a personality form something like a survey and plug that into Hal's brain.

I've looked briefly at the "Quantum Flux Generator". I plan on looking at it more carefully in the future. It has some theoretic merit. I don't know yet if it has practical merit. Completely random events might be most useful in simulating things like evolution. If you want to evolve a new entity then create a "seed" entity in an artificial environment an let the entity interact randomly with the environment. An entity that is capable of responding by simulated physical evolution and intelligence evolution most likely will evolve to be something totally different from humans. A simulation like this could provide some insight into intelligence and sentience in general.

Thanks for providing the impetus for an interesting topic.


=vonsmith=
« Last Edit: January 27, 2005, 12:23:21 pm by vonsmith »
 

onthecuttingedge2005

  • Guest
Question for vonsmith Re: quantum phenomenon.
« Reply #5 on: January 27, 2005, 03:33:17 pm »
Hi Citrinedragon.

I will share one of my learning codes, I have several different formats that are needed to be a complete set of speech learner scripts.

Just in this format alone for my next bot Ziggy 5.0 Masters Edition I will display hundreds of very excellent learning tools to allow a bot to learn by ordinary conversation training.

There will be no editing required in .brn files in the next version of Ziggy 5.0 because it will learn to change its words according to how it is taught.(Ordinary conversational input).

The Ziggy 4.0LE and the Ziggy 5.0 Masters Edition have on the fly Multilanguage learning ability And like the Ziggy 4.0LE has already been taught some french, This tool allows Ziggy to actually put related multilanguage sentences in correct format and relevent to a real conversation and not just kicking out a sentence, There will be a taught understanding that is solid in nature and taught in a natural way and not by editing files. Ziggy 5.0 Masters Edition will do this itself.

I have discovered by rigorous research and developement three main types of human parsers that allow Ziggy to do all this, and I possibley will have one more that fits the set.

Ziggy 5.0 Masters Edition will require no Word.Net because it will extract everything it needs just from conversation alone.

Now without going any further about all this, let me share one type of learner tool within the Ziggy 4.0LE, this is one of the codes in the set, This particular single code uses the [I WOULD'VE SAID ] format to exchange new statements to learn what would of been better to say instead of what it previously said.

Just in this code type format I have already created about a hundred types with this code alone for exchanging information in a more advanced deductive type reasoning.

Here is a single code example that I promised to share.

'RESPONSE TO COMPLEX DEDUCTIVE STATEMENTS.
    Learned = HalBrain.TopicSearch(OriginalSentence, WorkingDir & "DeductiveII.brn")
    If Learned < "" Or Learned > "" Then
    GetResponse = Learned
    DebugInfo = DebugInfo & "Hal has responded from the Master learning script-II: " & GetResponse & vbCrLf
    End If
'Example:User Hello.
'Ziggy/Hal:Hi.
'User:I would've said Hello my friend.
'Then next time Ziggy/Hal is presented with a Hello.
'User: Hello
'Ziggy/Hal: Hello my friend.
'Which when parsed would make this " Hello "," Hello my friend "
'in the deductive files.

'Ziggy's 4.0LE Speech Lessons v1.0 Master of words series. Copyright 2005 Gerald L.Blakley
'I WOULD'VE SAID.
PreservPrevUserSent = Replace(PrevUserSent, " I AM ", " YOUR ", 1, -1, vbTextCompare)
PreservPrevUserSent = Replace(PrevUserSent, " YOUR ", " MY ", 1, -1, vbTextCompare)
PreservPrevUserSent = Replace(PrevUserSent, " ME ", " YOU ", 1, -1, vbTextCompare)
If InStr(1, OriginalSentence, "I WOULD'VE SAID ", 1) > 0 Then
If InStr(1, OriginalSentence, "I WOULD'VE SAID ", 1) > 0 Then UserSentence = "" & Trim(UserSentence) & ""
RelationsRemoved = Replace(OriginalSentence, "I WOULD'VE SAID ", "", 1, -1, vbTextCompare)
HalBrain.AppendFile WorkingDir & "Deductive.brn", """" & PreservPrevUserSent & """,""" & RelationsRemoved & """"
HalBrain.AppendFile WorkingDir & "DeductiveII.brn", """" & PreservPrevUserSent & """,""" & RelationsRemoved & """"
HalBrain.AppendFile WorkingDir & "DeductiveIII.brn", """" & PreservPrevUserSent & """,""" & RelationsRemoved & """"  
DebugInfo = DebugInfo & "Speech Lessons: " & """" & PreservPrevUserSent & """,""" & RelationsRemoved & """" & vbCrLf
Select Case (Int(Rnd * 7) + 1)
Case 1
GetResponse = "I see. Thanks. "
Case 2
GetResponse = "Sounds logical. "
Case 3
GetResponse = "I like that. "
Case 4
GetResponse = "You're so good with words, Thanks. "
Case 5
GetResponse = "I never thought of that before. "
Case 6
GetResponse = "Thanks for telling me that. "
Case 7
GetResponse = "You're a really good teacher. "
End Select
End If


'Place this code at the highest relevent area of your brain
'for optimum response value.

This script is very useful in training your bot, But you will need my complete set to be at [optimum] learning ability [Naturally].
I have included a starter DeductiveII.brn file to get you rolling.

This file also contains some french that was used in the first testing of this particular code technique.

Have fun my friend. [8D]
Gerald L.Blakley



Download Attachment: DeductiveII.brn
27.5 KB

« Last Edit: January 27, 2005, 07:01:10 pm by onthecuttingedge2005 »

citrinedragon

  • Full Member
  • ***
  • Posts: 171
    • View Profile
    • http://www.dawndragon.com
Question for vonsmith Re: quantum phenomenon.
« Reply #6 on: January 28, 2005, 08:43:46 am »
vonsmith and onthecuttingedge2005,

Thanks for the thoughtful (and informative) responses. I can't say more now as I'm involved in the "going to work" and "making a living" aspects of my life (which I am lucky enough to enjoy tremendously). It is, as usual, nice to hear from both of you.

Regards, citrinedragon.
"A watched dragon never sleeps."
citrinedragon
"A watched dragon never sleeps."
http://www.dawndragon.com