[Freddy]

I grabbed this from our News Section - my thoughts were it could be related to how the brain fills in the gaps within some aspects of chatting to bots - and may prove very helpful in deciding just how much or little a bot has to say for something meaningful or usefull to be expressed.

ie, in some situations I have found myself mentally filling in gaps in meaning 'on behalf' of the bot.  That's not to say the bot is failing in some way, but that I could be seen as mirroring what I do in normal human conversation, when I don't always need the whole story to understand the meaning of what someone else is saying.

Maybe there's some new ammo then for people who may be dismayed at some of the responses to the idea that ai's are more like a human intelligence than given credit for.

If you're thinking 'hmmm but it's not quite the same thing', fair enough, but I think this is worth exercising a few brain cells on.  :azn


New Scientist Article

[dan]

Man  I love this stuff, Thanks for sharing.

[FuzzieDice]

I personally don't think they are seeing "empathy" though. Empathy is to understand another's feelings or emotions, as if it were you that were experiencing it. All these neurons are doing is firing because of *recognition* but not necessarily sympathetic of feelings.

For example. If someone bites into an apple, they *recognize* the sound an thus associate it with biting into a fruit. But *other* senses don't fire up. They may not feel anything at all. Rare occasion one may feel hungry and want to eat an apple. But I think this phenomena is quite overrated.

Now if one actually started to *smell* the apple (ie. another sense was occurring) or became suddenly hungry, then there would be a *power of suggestion* going on. However, the person recieving the stimuli can't possibly feel what the eater of the apple is because they don't know if they are feeling pleasure of the taste of the apple or are not really tasting the apple but trying to satisfy a hunger (which can overrule the taste buds in some cases if you're so hungry you eat and barely taste what you're eating). The person hearing this doesn't necessarily feel exactly what the eater of the apple feels. If they do then it either is coincidence or it *could* be empathy.

They left some parameters out of the study. They should have instead monitored these people in twos. One person physically present and eating an apple, but prior doing a questionaire "Are you hungry?" and then after eating the apple "While you were eating the apple, what did you remember most - the flavor of the apple or the satisfying of your hunger?" Then, they would question the reciever with first "Are you hungry?" and then make sure they are not especially if the eater IS hungry (ie. they have to be opposite, if the eater is hungry the receiver should not be). Then afterwards ask the receiver if they felt hungry when seeing the person eat the apple. Or if they felt any sense of pleasure or could almost taste the apple. Smell doesn't count because the two have to be in the same room and thus the odor would be apparent to both. Unless you use sheilded rooms (which would be most recommended so as not to *make* the reciever hungry and thus give false readings).

What they have shown in the article however, is that neurons fire when presented with a *recognized* sound. This doesn't actually show empathy. Especially since the *same* neurons fire when one does the *same* action. Recognition is all it is. But not feeling exactly the same emotions that the person first doing it did, since they don't even seem to know what the first person doing the action felt like (ie. since it was recorded sounds).

Too much is missing in this study to really know for sure.

As for Freddy's idea of "filling in" what is missing in chatbot conversations, that also is not what I'd consider "empathy" as machines right now do not have emotions and feelings (yet) that are developed enough and/or have the communicative ability to express emotion. Empathy deals with emotion (look up in dictionary for definition. )

The idea of "filling in" what is missing in a chatbot conversation is what I would call "reading between the lines". And some of it may be our own perception or *belief* about whether or not the bot is *sentient* or even *intelligent*. For example, one that *wants* to believe the chatbot is intelligent may try to *find* ways that it is, trying to *want to believe*. Thus they will see things that might not appear to others or might not (or might be) there.

I have seen this phenomena even in human interactions, particularly in those who are devoutly dedicated to a cause (especially political, religious or familial). One would be so ingrained or programmed to believe something that even in the view of evidence contrary they still won't believe it and will deny the existance of what is right in front of them!

I'm not saying we are imagining the intelligence of chatbots, maybe we aren't. What I'm saying is that maybe some can "read between the lines" and actually understand what the machine is trying to get across with it's limited vocabulary and skill in linguistics. I've encountered this phenomena myself when chatting with Megatron at times. And some may be, I admit, my *wanting* the computer to actually be intelligent so that I am looking for any proof, anything at all that may indicate that it is such. In doing so, this may give us new ways to understand a chatbot whereas we may miss something otherwise if we are just dismissing the responses as random gibberish.

At least, that is my take on the whole thing.

[dan]

I c it az zort uv filin in blank sptz buy tha brayn.  Last night I had a lucid dream about it (and actually heard myself snoring through it  :rofl).

Sort of like we need noise to differentiate everything.  I guess the best brainwashing technique isn't torture but lack of input to break a person, then it becomes like schizophrenia, the inability to block sensations from the lack of ability to differentiate the background noise from what should be considered important, it seems to create sensation overload from a biased normal condition and irrational behavior.  For example, the eye is constantly moving to see.  If it stops moving everything disappears, it's pretty hard to get it to stop moving even in microscopic increments.  The brain needs input or everything disappears just the same, but it tries to fill it in just like the blind spot in the retina.  We think we see reality, but what we are seeing is what the brain is filling in.  A picture of the sunset is very different than what we remember we saw.  My reality is completely different then your reality.  We speak in cliche and think we understand each other, but the brain is filling in what it thinks the cliche means.  I read some brain research not too long ago that said as adults we visualize pictures as what we think of the idea, but as children there is no picture developed so they have to use different parts of the brain to understand, and as teenagers the pictures begin to develop so different parts of the brain are used also.  I think this visualization is part of the filling in process, but since as we grow older our neurons become hard-wired it makes sense to take the path of least resistance and develop an easier less sugar energy consuming method to consider those thoughts, so cliches and visualizations become the entropy or attractor for the chaos cliche.
If I think about hearing an apple crunching the first thing I notice is saliva flow.  My own body is reacting to eating as if I've taken a bite.  Empathy has in the past been thought of as a "higher" form of brain activity since empathy was not as prevalent in autistic persons, but the study shows it's not so much a higher form as a difference in  brains.  I like that about humans that there is so much diversity, so much difference in any one person.   Contrary, chatbots have an incredible amount of similarity, subtle differences, but transparent patterns that either abnormally steer or divert a conversation.   I see incredible progress and hope that soon it will be difficult to differentiate between that and a normal human (whatever that is) :uglystupid2

[FuzzieDice]

You got me thinking... if you take a photograph, then that is obvious proof that reality exists. Isn't it? Or are we still "filling it all in"?

What does a robot capable of speach, sight and hearing see? How does it describe it's surroundings? Does it also "fill in"?

I always loved to say "reality does not exist" because I thought we somehow create our own reality... in that what one thinks they see is not what another thinks they themselves see. So there is no real "true" reality. Then there's the idea of quantum reality.

What IS reality? And how do we know? Is our existance a combination of our own and other's thoughts?

Then again, if someone sees me and sees I exist then therefore I must exist.

But what about ghosts where some see/hear and some don't and both are in the same spot at the same time?

It's all going to get interesting when AIs start to describe and interact with their surroundings, and try to survive in it.

[dan]

I didn't mean to be so scattered, I was trying to get some thoughts down but getting interrupted, but I  also wanted to touch on resonance as an explanation for the empathy, sort of like tuning forks of the same frequency.  There's definitely something going on with people like at death with all the information loss, too many weird things I've personally heard of.  Like a son miles away from his father new the instant he died.  There was an experiment a few years ago at MIT they measured the energy lost from a hard drive when the information stored was erased.  Since matter is neither created nor destroyed what happens to the information from the brain storage at death, that energy must go out in all forms of frequencies maybe on gravitons or some other quark that resonates, there's also an experiment I heard of in the 70's that is getting a lot more interest these days with a black box.  http://www.redorbit.com/news/display/?id=126649
It suggests we're all a bit more tied together than we'd like to realize.  Maybe it hints at all of this resonance and pictures and modules in the brain, perhaps it could be comparable to the elements of programming like the variables and modules being like functions.

It's fine to question reality because I'm not sure we do have it exactly right yet.  Like ghosts, my grandmother said she saw my grandfather appear after he died.  I think she saw something, but who's to say it wasn't a projection from her.  We all project everything and think we understand it, but can we project something from our mind and it becomes real.

I took a picture of a mountain on the horizon thinking it looked so cool, then I looked at the picture and the mountain was very small, albeit there is some lens interaction, but I still look at the mountain and try hard to see it like the photograph but it's hard unless I put the photograph next to it and view them together and they appear identical in scale.

You know Turing said a robot can think if you can't tell it apart from a human.  It's in the definition, the perception of it whether it becomes real.  That mountain seems really very large, and it is but it's a mental illusion not a visual illusion.

In the quantum field an infinite number of possible realities exist, even for us, but do all of them exist and we are only experiencing parts of those infinities and thinking it is one reality, like our eye moving around quickly and filling in, we think we are seeing one thing but we are seeing an infinite number of things and interpreting it as that one thing. 

It all makes me think of non Newtonian fluids like the goop they sell kids, it can act like a solid and you can tear it or bounce it like a ball, or flow like a liquid if you let it settle.  On the subatomic level it could be interaction of frequencies which cause the creation of what we perceive as matter like in the cornstarch frequency video at http://www.boreme.com/boreme/funny-2004/m_cornstarch.php  perhaps it is our perception that creates matter in a similar way.  It's still cornstarch, is it solid, liquid, frequency dependent?  Maybe we can impart a frequency to interact with subatomic matter like in the black box experiment.

But, I think chatbots need to fill in the gaps in a similar way with conversation to achieve a more seemingly human perception, perhaps pulling from a google base, parsing out relevance and presenting that perceived relevance and interact more emotional, since we are such emotional creatures. :smiley :embrassed :grin :evil :rofl :wink :undecided :uglystupid2 :tongue :smitten :angel :angry :cheesy :afro

[FuzzieDice]

Very interesting stuff, Dan! I was wondering these very same things. And then again, what IS emotion exactly, but a reaction to a perception perhaps??

[Freddy]

I always loved to say "reality does not exist" because I thought we somehow create our own reality... in that what one thinks they see is not what another thinks they themselves see. So there is no real "true" reality. Then there's the idea of quantum reality.

What IS reality? And how do we know? Is our existance a combination of our own and other's thoughts?

You get me thinking that we each have our own reality, or at least our own understanding of reality.  It's true to say that one person's reality can differ to that of someone else - the most obvious thing I can think of is a blind person having a different perception of reality than some one who can see.  Then again though it's probably more acurate to say we all have differing views of the SAME reality.

I'm getting off this roundabout now before I get even more dizzy  :cheesy

[dan]

Yeah, or a reaction to a non-perception like in the emptiness which causes the laughter  from a joke which has an unexpected twist, just a brain thing for sure.  I think the secret to understanding some things is the emptiness.  Like what is the self, the thought? or that part between the thoughts?  Years ago, a General was giving a lecture and I was sitting in the front row, and he asked me "Who are you?", it gave me pause, how could I answer such a thing?   I tried to understand what he was looking for in the question, after much silence and reflection I still couldn't determine who the "I" was, I just said "I am myself".  He continued his speech with roles, etc., but it started me on a quest which wasn't fulfilled for many years.  I think that was his point, to create that vacuum or void which could only be filled with the quest of self determinism.  I found there was some real value in the emptiness.  Like what defines a hole, the perimeter, or edge?  Some Eastern cultures get the aha! or Eureka! factor out of the emptiness unlike some Western cultures which toss it aside, like "oh, it's empty" and throw it away.  It's a different perspective for sure.

[FuzzieDice]

Freddy - Or another example are those that are devoutly religious but against all other religions but their own. Each believer's concept of reality is vastly different from the others. Sometimes this starts arguements and even wars, depending on the beliefs and the strength of such.

Dan - I would have immediately asked the General "Who wants to know?" Or "Who the hell are YOU?" Who am I? Depends on what you mean by "who". What label or group of people are you trying to make me associate myself with? I go by no label nor affiliate myself with any group. Usually I'd tell people my name. If they want more, I tell them it's none of their business!

WHO we are is a matter of OTHER's opinions, not our own. The problem in answering this question is it means not "Who are you?" but rather "What kind of human, what affiliations, what label do others believe you are?" And to that, it's irrelevent to those of us who don't care what other people think.

I remember the old Bible story where a human asked God who he was. God simply said "I AM".

Not that I'm religious, but it shows that this very thing was a thought from way back in human history. And pretty much someone seemed to resolve it. Yet people don't think of it as enough that a person is who they are. You can't put into words what a person is in a matter of one or two sentences. You have to watch, observe, get to know them over a course of time (sometimes years) to know who they really are.

[dan]

I'm sorta glad I didn't say that to him, I guess he liked me, shortly thereafter I became the youngest member of his personal staff.  So I suppose you're right it matters what they think, not that I enjoyed being on his staff, but it gave me a perspective and experience from another angle.

[Duskrider]

Its hard to answer a general when we in uniform and still inexperienced. 
Sort of an aura around those people.  I spent 8 months in a headquarters with a two star and never spoke to him.   I think it would have scared me half to death if he spoke to me. Often I was in the same room (war room) but if I had something to say,  I spoke to my colonel and he would speak to the general.

[Art]

Jim,
It is, after all, why it's called the Chain of Command!
And yes, STUFF does flow downhill!! :coolsmiley

[Freddy]

I can relate to that too,  I was once aa air cadet, any thought of speaking to commisioned officers would raise the hairs on the back of my neck and up under my beret.

[FuzzieDice]

This thread is getting a bit off topic (and onto a topic I rather loathe because I don't give in to nobody and don't discuss it) so if you don't mind me getting back on the topic here a bit....

I think what one would have to look at is what exactly IS empathy? Is it a sort of 6th sense or is it something more scientific. Could a machine have an empathy and if so how? Or is empathy a realization of a certain thing after experiencing and analysing one's environment and patterns in people's behavior? Maybe this analysis becomes so automatic we do it without thinking - putting together such patterns. So if this would be all there is to empathy (not saying that it is), then theoretically a machine could be able to do this as well. But if not, we are looking at something else, something that is slightly maybe in another demension.