AI- after years of speculation

[andrewcellini]

Ophiolite, you're hilarious. That is exactly how I speak with my friends.

you repeatedly call each other big dog and guy even out of context? that's what your program currently does.

 

i think it's more of a low end chat bot than an intelligent agent. maybe with improvement this will change.

 

how are you programming this? what is the code?

Edited October 3, 2014 by andrewcellini

[physica]

popcorn, you were making it out to be revolutionary and amazing. In your status updates you were hinting at the Turing test. At the start of this post your were going on about how you came up with the concept when you were a kid as if you were a child genius. The average smart phone has better comprehension of what I've said compared to your program. This is similar to when you were saying that philosophy grads can ace working in any field and when you were asking people on this forum to take you under their wing for graduate study.

 

The problem with you is that you know so little that you think that your half-baked amateur attempts at things are cutting edge. Then you think you're extremely smart because it's taken you a couple of weeks/months to bash whatever it is you've done together as opposed to real professionals and scientists who take years developing something. If you spent the amount of time you've spent trying to trick people into thinking that you're really smart into actually learning something, you may come up with something cutting edge one day. The problem with this is that it takes time and patience. You will get stuff wrong and it's not a huge ego massage to be learning something that loads of people already know. You may just see this as a personal attack but I'm trying to get you to stop wasting your time and actually do something constructive. No one here is buying it, your huge negative rep should have already told you that no one here is buys your narcissistic waffle. You have to ask yourself, are you actually interested in science or are you simply using this as a vehicle to get praise and attention? If you're truly interested in science then use your time effectively and actually learn what's already known. This forum is amazing and will help you along the way. There are some really clever people here who will help you no matter how mundane your question is. It's irritating to see you waste their time on nonsense.

Edited October 3, 2014 by physica

[Popcorn Sutton]

Yea we call each other big dog, guy man, budzo, buddy boy, gale, junior, standards, hoss, big hoss, big guy, chotch, and others too. Me and my friends don't judge each other. We make weird noises and say what's on our minds but rarely, if ever, do things get awkward and almost everyone that we encounter likes us.

 

I don't know what I did to it. Here's something that chomsky says to which I can relate. He says "in the beginning of the semester we usually start out by saying 'ok, let's assume this is true', and then by the end everything just falls apart."

 

The good thing is that because I've rewritten the program over a hundred times I have a pretty good idea of what absolutely needs to be in there for the program to be functional. I wrote a really good copy a while back which I showed to my boss and my friend which really impressed my boss and had my friend laughing pretty hard. Come to think of it, I was confident that the thing didn't need any more work, but at the time I was worried about efficiency and I think that I tried to solve that problem the wrong way which led to a path of revision after revision. I'm going into work today to grab my computer so I can go to the park, have a few drinks, and pump out that extra little bit of code to solve the issue that I had at the time. I think you guys might like to hear my argument as to why I have it set up the way I do, so I'm going to take the time to write it out for you guys.

 

I wrote the program in python and it's going to be up on the internet indefinitely pretty soon, but I'm going to have to work on the other aspects after I make one minor revision to the one that I had (and one major revision to make it more unitary, but I'm probably not going to worry about that for a bit because of how much time I've spent on it already).

 

Anyways, here's the argument.

 

You are you, I am me. We never confuse ourselves with each other (plainly). When we meet, you have little, if any, disposition towards me, and so you use your fall back data- knowledge. Once we've built some common ground, if we ever encounter a particular context that has a length less than two, if the length is one and the only unit in context is the unit that is being contemplated, you lose your train of thought and, if there is no match to the unit being contemplated in disposition or knowledge, then you fall back on your most similar and most recent match in your disposition towards the interlocutor, which could be within context, but most of the time it just seems that we start talking about other things. Input inherently has order, so do not mess with that order. For every unit in the ordered context, if the strength of the unit is stronger than the previous, add it to the output but also add the strength to another set to help eliminate unwanted units (this works wonders). For every unit added to the output, move that unit to the most recent position in context (there's evidence for this linguistically such as displacement and long distance agreement).

 

I don't like for loops so I've made it a point to avoid those, but after the insight I had last night I've changed my mind.

 

I hope I'm not missing anything here. Once again, thanks guys, you've been great.

Edited October 3, 2014 by Popcorn Sutton

[Popcorn Sutton]

(Mods- I don't know why you guys continue to let physica act this way. It's seriously disturbing.)

[Popcorn Sutton]

It says "what" in response to every input.And I have given you a downvote to get you to the round -100

That was the nicest down vote I've ever received haha. Thanks buddy!

 

The equation I have for knowledge acquisition is ready for the archives by the way. I shouldn't be doing this but here it is.

 

u = '(o)'*n

r = (u:{strength:n, individuator:1 or 2, order:[], r*n})

t = {order:[], r*n}

m = [(P(u|t[order]))*n]

t = (r[m])*n

 

u = unit of knowledge

o = occurrence

n = any positive whole number including zero

r = self referring recursive structure element consisting of order, strength, separation of minds(individuator), and itself.

t = time (consisting of order and structure; knowledge)

m = mind (a set consisting of a progression of units that are more likely than the previous)

P(...) = probability of (...)

| = given

* = multiplied by

: = defined as

[] = list

'' = string

{} = dictionary

 

My argument is that this is the equation to encapsulate the mind in relation to the structure of the brain.

Edited October 3, 2014 by Popcorn Sutton

[andrewcellini]

Yea we call each other big dog, guy man, budzo, buddy boy, gale, junior, standards, hoss, big hoss, big guy, chotch, and others too. Me and my friends don't judge each other. We make weird noises and say what's on our minds but rarely, if ever, do things get awkward and almost everyone that we encounter likes us.

it's not that there's anything wrong with what you say or have the program say, it's the repetition of these phrases without provocation that makes this program not so fun to interact with. if i for example said "what's up?" and it said "not much buddy" there would be no problem. it shouldn't say "what's up buddy" for any given input.

 

I don't know what I did to it. Here's something that chomsky says to which I can relate. He says "in the beginning of the semester we usually start out by saying 'ok, let's assume this is true', and then by the end everything just feel apart."

 

The good thing is that because I've rewritten the program over a hundred times I have a pretty good idea of what absolutely needs to be in there for the program to be functional. I wrote a really good copy a while back which I showed to my boss and my friend which really impressed my boss and had my friend laughing pretty hard. Come to think of it, I was confident that the thing didn't need any more work, but at the time I was worried about efficiency and I think that I tried to solve that problem the wrong way which led to a path of revision after revision. I'm going into work today to grab my computer so I can go to the park, have a few drinks, and pump out that extra little but to solve the issue that I had at the time. I think you guys might like to hear my argument as to why I have it set up the way I do, so I'm going to take the time to write it out for you guys.

do you have the original code to this version? did it work in a similar way to your current version?

Edited October 3, 2014 by andrewcellini

[Popcorn Sutton]

do you have the original code to this version? did it work in a similar way to your current version?

Yes I do, it was written on September 9th and I will be adding the two necessary components that I just didn't think about at the time. Namely, statistical deletion of units that don't occur within the typical knowledge ratio of the context as well as statistical inference of unknowns within a proximal domain.

[imatfaal]

!

Moderator Note

 

Guys and Gals - can we tone down the personal criticism? Please feel free to castigate the idea - but no more cod-psychology, veiled (and not so veiled) insults, and comments on other members and their characteristics/personality.

 

Thanks - do not respond to this moderation

 

[andrewcellini]

Yes I do, it was written on September 9th and I will be adding the two necessary components that I just didn't think about at the time. Namely, statistical deletion of units that don't occur within the typical knowledge ratio of the context as well as statistical inference of unknowns within a proximal domain.

how does this program "know" or "learn" from input to input?

 

how does it mimic?

 

edit: i guess i should really be asking what is the algorithm for mimicry?

Edited October 3, 2014 by andrewcellini

[Popcorn Sutton]

It mimics only when the output (often without vocalization) is equal to memory. The program learns based on the output (again, not always vocalized). The point of interest (or the unit of contemplation when it is not "split" unitary, split unitary meaning that it's not a "maximal unit" which is the entire input).

[Popcorn Sutton]

Memory is determined like this.

 

If the length of the output > 0, memory equals output.

 

Pretty simple operation. Vocalize the output of the length is not equal to zero, else output equals memory. Don't vocalize memory because it's a duplicate.