And we're off

[Ali Helmy]

Hey mates,

First of all, let me just say welcome aboard and onto the Artificial Intelligence Society

Secondly, i'd like to say that I'm quite impressed with the response for the society, as we have currently 15 members, and that was amassed in just two days. I hope that this enthusiasm will accompany us along the way

Thirdly, on to business, we need to simply keda try and define or more vaguely just describe what we want to do as our next task. Is there something anyone wants to share? an article? a new technology? anything would be most welcome, even just a philosophycal discussion regarding AI, and every now and then, some humour cant hurt

In all, i hope we all enjoy and benefit of this society as we can, and that it shall help us in any means possible. in the mean time, I will leave you with this link, from the AI: a Modern Approach [AIMA] book, with HUNDREDS of useful links

--
Cheers,
     Seth
     <A. Helmy>
==================
  One Life... LIVE It
==================

[Sherif Elian]

First of All , I think this society is a great idea  , I don’t know a lot about AI yet , but I like the concept of MAKING intelligence , not just using it. Also I like the topic because it is very philosophical and I like philosophy alot.

 

As this society is dead since it is up , I thought to share an idea with you all , maybe it is not interesting a lot as a start but this is all I have in mind now.

 

I was thinking from sometime about not only make the computer thinks but also make it feels  and I was trying to answer this questions

-is it possible to make a machine feels and if so , which level can we reach , and what kind of feelings can we make

-is it useful  and if it is , are all the feelings useful ( I believe fear is useful in human being , is it in machine ? ) (what about hate ?).

-is the feeling different from thinking  or it is kind of, how we feels ?

-why human feelings is not the same ? does it depends on the character ?

-can we add a character to a machine ?

 

Just think about it , very interesting.

Sherif Elian

[Mohamed Ali]

Sherif,

Thanks for this nice startup. Let's move on,

I think we shall need to answer the questions of : What, Why and How. I don't know in which order we shall answer them but let's move,

First What is feeling, Can it be a function in character, mood, and current perception!?. Are feeling and thinking different or there is somehow a relationship between both of them leading to persons' actions as an output!?. Can we say that in many cases thinking is affected by feelings as an input? But can feelings be affected by thinking?

Second Why feeling, I mean why for a machine. At this point the objective must be clear. Can we say that this can lead to a new generation of interactive applications moving towards the engineering approach of making people's life better, easier and more interesting in a scientific, economical way?

Third how to make feeling, Currently I don't know!!!, But moving forward in the AI science can give us a better view for that.

I don't have a proof of what I'm saying but it is all Feeling :).

Thanks again for this nice beginning, let's share thoughts and feelings to enhance our way of thinking.

Regards
Mohamed Ali

--
--------------------------------------
Mohamed Mohamed Ali
Computer Engineering Department
Faculty Of Engineering
Cairo University

[Ali Helmy]

Ok lads, I have some comments of my own here...

Well, first of all, but can everyone please stop top-posting, ie: when you reply to an email, place ur reply AFTER the quoted text, so the message history can easily be traced and understood...

Now, onto the GOOD stuff :):)

I really dont understand human feelings anymore than the next bloke, but what i know, and what Neurology would tell you is that it is 'no more than a sequence of chemical reactions ', but then again, why the hell do we get upset when we cry? So, i really dont know how far we can implement them, because i doubt we have a good graps of understanding to start with... So the WHAT and the HOW i completely have to admit, i leave blank...

Now, onto the more interesting point... the WHY... you [sherif] said that "I believe fear is useful in human being , is it in machine ? ) (what about hate ?)", well, i have to admit i disagree... i mean, pain is important for human beings, it tells us when something is wrong, and scientifically, it IS a feeling, but i know thats not what u were talking bout... i mean FEAR, HATE, LOVE, and other i think are for the sake of larger and better communities, motivation, etc... but do we really need that in a computer? I have to quote here a few words from the AIMA book, which may further explain my thoughts:

'Yet researchers have devoted little effort to passing the Turing Test [it evaluates how much a computer can 'act' like a human], realising it is more important to study the underlying principles of intelligence than to duplicate an exemplar. The quest for 'artificial flight' succeeded when the Wright Brothers and others stopped imitating birds and learned about aerodynamics'

Why in my opinion? because creating a plane that looks, feels and flies like a bird is not usefull... so creating a computer that can talk, feel and fo shopping like a human is not useful!

Interesting topic though, touche mate

[Dooma®]

I agree with Ali that we have to discuss first the motive for having such a thing.
in my vision of the topic, I think that feelings are useful when it affect the machine's descision.
Think you have two tasks, one of them to serve a user that treats you kindly, and another who treats you bad, and you have to choose which of them to serve first, or which of them to serve. I think this concept is already used for example when the OS gives priority to a task that is known that it don't cause troubles.

But what about Sadness. How sadness can affect the machine's desicsion? I think it is useless unless you are designing a software friend, and don't want him to start laughing when you tell him that you lost your job. But even this can be done, it can be a mode or a state that the software can go into when seeing you depressed.

Do the machine have to regret something? what will it benefit from regret anyway? maybe for future learning? Isn't this concept already used for machine learning?

what about fear ? How would fear affect the machine's decision?

I can't tell wither feeling and thinking are the same for human beings, but it can be always translated to thinking in case of machines.

In brief, i think that all emotions than can be though of as a way to change the machine's decision can be implemented, and other emotions that have no impact on the machine's decision are useless

feelings can be implemented even if we know noting about how human feels. We know very little about how we think. However, AI has emerged

[Mohamed Ali]

Well,

I think it seems irrational to discuss the Why before the What. How can we discuss the advantages/disadvantages of something we don't agree to a definition to, even an abstract definition?

This definition of feeling can be a beginning for answering the What question. Then we can move towards the Why.

Regards
Mohamed Ali

[Ali Helmy]

Good point Mohamed, and i must say , here i have to quote wikipedia's definition for simplicity:

Feelings are affective states of consciousness. They are typically triggered in biological beings by information in the physical world and, in the case of humans and complex animals, thoughts. Note, however, that feelings can arise even in the absence of physical information.

Now, i have some lines to draw on that...

A) States of conciousness... did i read that right? i tink i did, and i think that is a very useful thing that we can all relate to very easily... STATES... how many times have u had to design a system or code that had different 'states' in it to the approach? maybe that is really all we need to imply to generate machine FEELINGS... i'm saying we're not going to develop machines that will laugh and joke, but will assess the situation according to their STATE of mind...

B) typically triggered in biological beings by information... ie, thoughts... now i think this is also very interesting... so following the model i was just talking bout, these states of conciousness [or you may think of them as different code paths to be executed] are determined and generated by the information gathered from the AI agent's environemnt... now this would truley build up on the use of these 'machine feelings' and would thus start to present itself in a more 'engineering' sense, as apposed to just dreams...

C) feelings can arise even in the absence of physical information... finally, the very peculiar point of perception... why DO we get affraid? is it because we have a bad memory that when we [or someone else] did 'this' they got 'hurt'? or is it simply that we do NOT know of anyone [us or anyone else] doing this before, and we do not know if we would get hurt? now these two concepts again put us in a situation where the system would benefit from 'machine fear'... in cases where executing a certain instruction or action [program, etc] have led to previous mishaps and damage to the system, perhaps the AIAgent should be more careful... similarily, it would be quite useful to be careful the first few times u run an un-signed program for example [firewalls, anyone?] just to make sure ud get to know it first...

So, what im saying is... if we take by the Wikipedia's definition, then emotions and feelings are somehow a sort of 'thought state' presenting different branches on the 'decision tree' for the party involved... in that sense, it would be very beneficial to develop machine feelings, to help correctly shape and evolve the machine code...

But, i pray to you, would u ever imagine use of having your computer weep after playing a sad movie..?

[DoomaR]

Sorry Mohammad, I thought that we all agree on the meaning 8-), but u r right