Week 3 [16.11-22.11] Can machines read our emotions?
Computer systems and machines are becoming the foundation in every domain of the human life. Up until now there has been a strict, practical division between what is the human and what is the machine domain and in a way the two stood in contradiction to each other. Human means emotional, irrational, intuitive, social. Machine-like, on contrary, is purely logical. But is it really?
Computer engineers and scientists are looking for ways to incorporate human features into machines, especially in the domain of robotics. I am very curious to know where are the limits for this. In the Human Centered Multimedia laboratory a group of designers is simulating human-like behaviour of robotic agents, exploring how these can understand the social signs and emotions and react to them just like a human would do. They have created eight
different expressional designs for the emotions of Anger, Sadness,
Fear and Joy - these emotions are separated into Body Movement, Sound and Eye Color ans, subsequently, incorporated into robotic design and tested against humans.
Please read the article:
https://www.aldebaran.com/sites/aldebaran/files/casestudy_hri_augsburg_0.pdf
and answer the following questions:
1. Are human emotions really that simple that they can be put apart into logical elements and easily learned by a machine?
2. Can we do the same with our intuition? What could the design of intuition look like?
3. If our "human components" are that easy to put into logical elements, how are we different form machines?
This topic, like anything else, cannot be assessed with yes or no answer.
ReplyDelete1. There are areas when even machine would be probably pretty efficient in learning and mimicking human emotions. But there are at least few, machine would struggle or fail badly. I'd like to see machine trying to interpret a human sarcasm..
2. Design of intuition? This could be easily straightforward answer like.. coin toss. In some situation or hunch would be probably as probable as coin toss, right? So why not to simplify design?
But AI is being developed right now. And frankly speaking, no, I don't like that idea. On the other side it looks inevitable, especially considering Google having all the data access and capabilities to build it.. have you heard about TensorFlow? Latest news from Google? They released their Google Brain library. Machine intelligence is going to be increased by this move definitely as more and more date and usage of data will be provided. Will see were it will lead humanity. Elon Musk (SpaceX and Tesla) called the prospect of AI 'our greatest existential threat'.
3. Ironically, we're not in fact. Each of us has its own set of routines. How that distinguish us from machines?
Thank you for your comments.
ReplyDeleteYou are right, sarcasm employs not only a bitter, sharp remark, but also ambivalence. Hopefully we still shall be able to refer to this emotion as 100% human, or shall I say "impossible to determine through entropy....". Maybe there is some other algorithm that could determine ambivalence?
As for TensorFlow - this is very interesting, I will take a closer look, I am really interested how intelligence can be built entirely on data. How are they constructing knowledge? (data is obviously not equal to knowledge)
I think that emotionality is very complex and any logical system is not able to describe it. For this to be feasible, we would have to create a logical system corresponding to human reasoning. The logic of two or three valuable this does not guarantee fuzzy logic - I do not know. In addition, each of us is different and the machine would have to be universal - I think that here is the biggest problem. If we are talking about intuition as on inference from the evidence, I think the machines are able to overtake us humans, both in time and in the correctness of reasoning. The problem was that just putting all the conditions ...
ReplyDeleteFrom what I understand you are saying that intuition, or human reasoning in general can be described as inference from evidence (motives, purpose, intentions, what else?) - which would be easy to compute. However you are also stating that everybody is different, so these unique conditions in which the inference occur are each time different and difficult to predict. Maybe it's a matter of grouping us, people, into a limited number of types, hence getting a limited number of conditions?
Delete1. Are human emotions really that simple that they can be put apart into logical elements and easily learned by a machine?
ReplyDeleteAccording to this article this is not very easy thing.I think that in some parts of understanding human emotions, computers and robots and can amaze us. For example in face mimics, when person gives to the robot very clear mimic expression it can interpret it as it should.
So in very obvious cases it can be quite easy but when our mimics and behavior becomes more complex it can be almost impossible to detect human emotions. But situation can change in few years.
2. Can we do the same with our intuition? What could the design of intuition look like?
It depends what kind of intuition you had on your mind. There are three main types of intuiton: physical, emotional and intellectual. And according to different religion and philosophy (in Psychology also - Freud and Jung) it's not allways the same :-).
But if we talking about nowadays description of this word (which is ofter misinterpreted as instinct, truth etc) as a "gut feeling" the answer is no. This is because computer
This is because computers have to have self-awerness which is also topic from psychology. Maybe in very far future this can be achieved but not now and not in the near future.
3. If our "human components" are that easy to put into logical elements, how are we different form machines?
Some physical parts of human body can be built from from logical parts (leg or hand), others won't . So in the same way humans parts can be similar to machine parts and sometimes they can be replaced.
Thank you for some really interesting insights from psychology. I didn't know there are three types of intuition - what I had in mind was intuition as reasoning / decision-making process that happens in our subconsciousness, and what our conscious receives as a result of the process is a simple insight (perceived as gut feeling, a thought, an idea, a voice....) that tells us "this may actually be like this..."
DeleteMimicking intuition, as you defined it, seems extremely hard. I assume that researches would need to first figure out what exactly is happening in our subconscious, and we’re pretty far from this. On the other hand, when it comes to making predictions and decisions based on historical data, AI/Machine Learning techniques are already doing this pretty well.
DeleteWhen it comes to emotions (question 1), I agree with Tomasz. There are certain classes of emotions which can be easily detected, but this is not the case for all of them. Often people in complex situations are not able to easily name the emotions which they are feeling.
I don’t believe that it’s easy to put “human components” into logical elements (question 3). I think that scientists are moving forward in emulating some parts of human behaviours and abilities, but putting this all together to build an artificial human sounds like science fiction. I believe that specialised machines/algorithms will be taking over some of human tasks over the next years but it
won’t be possible to emulate creative behaviour and abstract thinking.
You are right that simply emulating some "human parts", replacing them with artificial intelligence and then putting them again together will not give us a human being.
DeleteHuman emotions cannot be learned by a machine.
ReplyDelete"We imagine a future where robots will be in our homes to help us set the table, put away the dishes, motivate elders tinker ... This requires great intelligence and knowledge (including emotional). The robots must be able to perform complex reasoning "... Put the dishes in the dishwasher asks for example to know mobilize a lot of manual skill and intellectual. It is clear that to live with us robots will need intelligence.
But which one?
Human is perhaps the most complex machine in the universe
In humans there are two kinds of emotions.
The emotion of the reptile, our reptilian brain, the hypothalamus, which is limited to send in the body's hormones and neurotransmitters to stimulate hunger, reproduction, fear, etc.
And cognitive emotion produced by the brain cells, the emotions born from the experience, knowledge, which we are moved by an equation, a table, etc.
For the robot against him, how does he has emotion, the only intellectual emotion.
It has no hypotalamus but knowledge, and emotions born of knowledge, they are perfectly logical, cognitive, and he will.
So in terms of these emotions there, cognition, man and machine will be on an equal footing.
It is not a dogma, but a reality.
That's a very good point. And looking at what you are stating from the other side: I think that in the modern times many people are trying to consciously limit their feelings only to the intellectual emotions, to become this kind of logically thinking business people. I would say that the world is actually pressuring people to become like this, this kind of "intellectual emotional" thinking is said to be the means leading to success - would you agree?
Delete1. Are human emotions really that simple that they can be put apart into logical elements and easily learned by a machine?
ReplyDeleteI don't think so, compare our brain complication with even the fastest supercomputers. I'm not even mentioning how complicated it is to express our feelings.
Enjoy reading this article:
http://scienceblogs.com/developingintelligence/2007/03/27/why-the-brain-is-not-like-a-co/
2. Can we do the same with our intuition? What could the design of intuition look like?
Computers can predict, I totally agree but intuition is far more complex. Simple "example" based on Star Wars :)
C-3PO reported that R2-D2 predicted "the chances of survival are 725 to 1."
Still Han went out to find Luke :)
3. If our "human components" are that easy to put into logical elements, how are we different form machines?
Still as stated in my link, logic gates are far less complex than human components. We are not prepared to build as sophisticated pieces of hardware to emulate human components.
Thank you for an interesting article! You are saying we are not prepared to build sophisticated pieces of hardware to emulate human components, still, as it was mentioned earlier in the discussion - some reasoning tasks can be carried out much quicker and more efficiently by computers that humans, and given the speed of development in this domain - we may soon be prepared.
DeleteI don't think that detecting human emotions is "that simple". I followed the link provided in the article and watched a video "Emotional Machine Learning by Angelica Lim" on Aldebaran Robotics' channel on youtube. At 5:10 Lim show a diagram that decomposes emotion recognition to components like: facial expression, speech semantics, head angles, vocal emotion and few more. I think that current state-of-art machine learning systems can deal with these kinds of tasks. After classification algorithm has to put all the decomposed pieces together and that can be tricky. Sarcasm detector? Here you are: http://www.thesarcasmdetector.com/ (I didn't test it, I was just curious).
ReplyDeleteWhat is intuition, can we define functional requirements? Is it finding a solution subconsciously, based on previous knowledge/experience or simply a hunch, as previous commenter suggested? Maybe probabilistic generative model would generalize well?
We have much better CPUs than machines. Every neuron is a nonlinear computational unit with around 20-50 Hertz frequency, but we have a lot more of them than most powerful machines. And nobody has to develop (buggy) software for our computers, learning starts on its own with huge amount of training data - world around us. Brain doesn't have to deal feature extraction problem, it's out of the box. On the other hand DeepMind's reinforced learning algorithm beats human in arcade games.
I am definitely going to test the sarcasm detector :-) and another inspiring idea - but not yet developed - is the Manslator (Female language translator) - I bet a device like this would be a bestseller: https://www.youtube.com/watch?v=ezVib_giTFo
Delete1. Are human emotions really that simple that they can be put apart into logical elements and easily learned by a machine?
ReplyDeleteI don't think it can be easily learnt. For once, we tend to hide our emotions talking to some people and be more open with others.
And above all, there is an irony. Many people struggle to read subtle clues that conversation partner give indicating that he is not serious. Now, try to teach machine to understand it... It is next to impossible.
Of course, based on statistics (as usual) we can teach the machine the most probable interpretation of audiovisual clues. But... These clues differ based on ethnic origin and cultural background, as well as personality. We tend to unconsciously build mental models of people that surround us based on their past behaviour and actions. The mental model is obviously related to biases identified by Social Psychologists, namely cognitive bias and confirmation bias. Our own built-in neural networks continuously classify information we get from outside world and quite often it leads to prejudice and misreading other people's emotional state... Can (for instance) robots learn how to be good at reading emotions without going into details? Or using the subtlest clues? Personally, I don't think so. This topic is so complex that even we, social machines are notoriously poor at dealing with the subject... How can we teach machines to understand something that we... don't?
And to top it all, there are people with so-called personality disorders like Autism and Asperger's Syndrom. These people can give completely incoherent signals regarding their emotional state (raging from not showing any emotions while being touched deeply to overly expressing very subtle emotions). As of now, there is no single "normal" person that could understand them. If we want to build personal robots, or even conversation agents, we have to take these facts into account.
2. Can we do the same with our intuition? What could the design of intuition look like?
I believe this is extremely hard question. How do we define intuition? Going on with Wikipedia's definitions: "Intuition, a phenomenon of the mind, describes the ability to acquire knowledge without inference or the use of reason.", I can only say that we would have to teach machines to be biased. And I don't mean we have to teach them to understand confirmation bias (although it might be useful to certain degree), but to actively use confirmation bias. This begs the question: what is the "proper" confirmation bias? Should we teach machines to be for instance racists?
It's more the question about ethics not and less of the technical aspect of the process... Should we, or shouldn't we teach machines to use biases?
3. If our "human components" are that easy to put into logical elements, how are we different form machines?
I believe, I pointed out so many complexities above, that we really can't say that this is easy. Besides, the field of Social Psychology is pretty young (like 70 years) and we still have to learn a lot about ourselves. How can we even think that in a reasonable amount of time we can teach machines to become social actors, when we don't even understand social mechanisms to the last detail?
Can the machine in my lifetime somehow infer that that was rhetorical question?
We do hide our emotions on the conscious and verbal level, but on it's impossible to hide them on the non-verbal and subconscious level - I guess this is what makes the human-human interaction interesting. Just as discussed in the post below with Joanna - this mix of hidden and explicit, ambiguous or conflicting signal is what gives complexity to our interactions and gives incentives for the brain to keep engaging, insteead of instantly getting bored.
DeleteAnd thank you for the definition of intuition - in the previous post (conversation with Wiesław) we were wondering if we can refer to it as inference based on evidence - and now it is clear that we cannot
1. Are human emotions really that simple that they can be put apart into logical elements and easily learned by a machine?
ReplyDeleteAccording to my experience human emotions are not simple at all, but I am far from saying machines will never be able to “feel” emotions. Now we try to copy human gestures, voice or face expressions, but these is not real emotions. It is nevertheless true that coping with a humanoid robot might have an effect on our mood … until we get bored.
2. Can we do the same with our intuition? What could the design of intuition look like?
Some people say that experience is the source of human intuition. It seems nevertheless that there is something more than that. How to explain people dreaming the same dreams or avoiding death because of a strange coincidence ? A very simple design of intuition would base on the internet content, but it would be very imperfect. Maybe when people start to live more in the virtual world than in the real one, we will have it will be easier to predict their thoughts and behaviors.
3. If our "human components" are that easy to put into logical elements, how are we different form machines?
At the moment we are not even able to define the "human components". This is why I think there is a long way before we put into logical elements.
I think it's a good observation that a robot copying our expressions may influence our mood, but only to a certain point - until we get bored. That would mean we like and need the complexity of the emotional signals - interesting insight that I will use in my research, thank you
DeleteThis comment has been removed by the author.
ReplyDelete1. Are human emotions really that simple that they can be put apart into logical elements and easily learned by a machine?
ReplyDeleteIn my opinion human emitions is not simple and we can't put apart into logical elements and easily learned by a machine.Robots is computer and its help us about activities but no emotions.
2. Can we do the same with our intuition? What could the design of intuition look like?
In my opinion we can't do this with our intuition because people have a different intition.
. If our "human components" are that easy to put into logical elements, how are we different form machines?
I am agree with Joanna Sielewonowska . We are not even able to define "human components".
1. Are human emotions really that simple that they can be put apart into logical elements and easily learned by a machine?
ReplyDelete2. Can we do the same with our intuition? What could the design of intuition look like?
3. If our "human components" are that easy to put into logical elements, how are we different form machines?
There are some simple emotions like happiness, it is quite simple to recognize when someone is happy. That is why I am sure that machines can recognize some basic emotions. But there are also emotions based on context of the situation. Now that may be possible in future but currently it is impossible. Machine would have to understand the situation, understand the context of scene. This is very hard. The main difference between us and machines is that we can do something that is in opposite to a logical frame and the other person will still understand the true motive. Why ? Because we don’t only think using 0/1, we understand the situation, we understand the context. This is the biggest difference.
I am wondering if happiness is simple - some people don't know they are happy, waiting for something better to happen and always seeing that the grass is greener somewhere else - while the fact that they are safe, provided with all they need is enough to judge them as happy people. And on the other side you can be a person always laughing and bursting with energy but that behaviour could be hiding an empty whole in the soul
DeleteIn my opinion people who can't recognize if they are happy are exposed on some degree of depression and they can be adjusted with proper physiological or medical treatment.
DeleteI think that human emotions can be reproduce only to a point. Do not be reproduced in full, at least today.
ReplyDeleteIntuition like emotions can not be reproduced at 100%. But I am an optimist and I think it is a matter of time.
In conclusionI think that we can reproduce some of the thought processes.
What exactly makes you feel optimistic about the fact that soon human emotions can be 100% reproduced? :-) This idea makes me a little anxious, that's why I am asking
DeleteEssentially as was said at the PhD workshop there are many problems in this area: like implement in machine sentiment analysis, learn machine to summarizing text, and so on. I can't imagine different implementation method than that suggested by Michael. Its funny by the way. I also agree that people have their own set of procedures. Moreover, they have different experience, habits and points of view. People are very different from the machine.
ReplyDeleteInteresting approach to refer to human activities as set of procedures, unique experience and habits and point of view (deriving from experiences I would say)
DeleteReading this article I had a feeling that Steven Hawking spoke about this subject recently. I believe, it is only a question of time when machines will reach the human level of logical thinking ( and than surpass it furiously …. )
ReplyDeletehttp://www.bbc.com/news/technology-30290540
Thank you for the interesting link
Delete1. Are human emotions really that simple that they can be put apart into logical elements and easily learned by a machine?
ReplyDeleteIf we are talking about “simple emotions” like laugh, happiness, surprise, sadness then there is a lot of projects used mainly in marketing areas and are able to judge them quite well. The same in security systems which are looking for strange behaviors and can simulate next steps of it. Answer is - probably and partially yes.
2. Can we do the same with our intuition? What could the design of intuition look like?
Probably as colleagues told we are talking about AI. What is intuition? It is decision making process based on our experience, or just feeling (in most of programming languages - “random”).
3. If our "human components" are that easy to put into logical elements, how are we different form machines?
Thousands years of evolution.
Do you know any examples of such security systems looking for such abnormal behaviours? That would be interesting to my research
DeleteYes it is true that machines can learn human emotions. Human being express emotions through facial expressions by flexing specific groups of muscles. Those movements can be read by a computer equipped with a camera and some designated software.
ReplyDeleteBut this is not what we really mean by understanding, is it? When we say understand what we really mean is is a computer capable of comprehending human emotions and acting upon it. In that case, everything depends on the software. It is absolutely possible to create a program that would behave in certain ways when specific events are taking place giving the illusion of empathy.
Illusion of empathy - that's a good way to put it.
Delete1. Are human emotions really that simple that they can be put apart into logical elements and easily learned by a machine?
ReplyDeleteI don't think human emotions are simple to read, especially if you put cultural context into it. I guess Chinese, Arab and French guys are not reacting in the same way on the same stimulus. The same to projection of the emotions. I think NAO is good start but this is far away from giving the feeling that you companion authentic human emotions.
2. Can we do the same with our intuition? What could the design of intuition look like?
Many scientists claims that intuition is an output of background process that are running continuously out side our mind, so we don't have direct access into it. I think, because of that it is hard to understand how it works, how ever our mind experiences it as a result, as a projection of inference driven by our life experience. It might be possible that we can i.e. teach artificial neural network to act like processes which we commonly call intuition, but as we are never sure if our intuitions have right, how can we validate is it working correctly?
3. If our "human components" are that easy to put into logical elements, how are we different form machines?
I think we are still far away from understanding how our brain works. Can we already explain who our mind works ? Only when we understand the deeper mechanics that rules the entity we can have chance to mimic our self. Without that emotional imitation will look like NAO robots - like a kind of cartoon.
In my opinion intuition can't be approximated with artificial intelligence but I feel that we are thinking in similar direction.
Delete1. Are human emotions really that simple that they can be put apart into logical elements and easily learned by a machine?
ReplyDeleteThis topic is related to "Effective Communication" blog post because both deal with communication, this time machine-human one. I believe that emotions are very complex and we don't fully yet understand how they are formed. However I also believe that one day technology will mature enough to produce machines capable of understanding emotions.
2. Can we do the same with our intuition? What could the design of intuition look like?
I am not an expert in the field but for me intuition is like a process that runs in the background (our subconsciousness) and can automatically send impulses to our consciousness about future repercussions of our decisions (the brain automatically relates certain matching criteria to events that have already happened). I don't see why it couldn't be emulated on robots.
3. If our "human components" are that easy to put into logical elements, how are we different from machines?
We are not really different. We are made of organic material because this form was faster at producing a conscious being. It took millions of years of trial and error by random natural events and currently we are the final product. The topic is so broad (but also interesting) and already discussed in previous iterations of this blog that I was expecting it to appear in this one too. ;)
1. Are human emotions really that simple that they can be put apart into logical elements and easily learned by a machine?
ReplyDelete"Many psychologists believe that there are six main types of emotions, also called basic emotions. They are happiness, anger, fear, sadness, disgust, and surprise. Happiness is our reaction to the positive, as disgust is to the revolting and surprise is to the unexpected. Similarly, we react to aversion through anger, to danger through fear, and to difficulty or loss through sadness." I think that it is not so simple. Please have a look the following example. List of emotion I found:
(Acceptance, Affection, Aggression, Ambivalence, Apathy, Anxiety, Boredom, Compassion, Confusion, Contempt, Depression, Doubt, Ecstasy, Empathy, Envy, Embarrassment, Euphoria, Forgiveness, Frustration, Gratitude, Grief, Guilt, Hatred, Hope, Horror, Hostility, Homesickness, Hunger, Hysteria, Interest, Loneliness, Love, Paranoia, Pity, Pleasure, Pride, Rage, Regret, Remorse, Shame, Suffering, Sympathy) How can someone think that it will be ease to emulate only part of them ?
2. Can we do the same with our intuition? What could the design of intuition look like?
"Intuition, a phenomenon of the mind, describes the ability to acquire knowledge without inference or the use of reason." In my opinion that " phenomenon of the mind" was build during human evolution. It is 200,000 years old. Nowadays we can try to emulate part of intuition behaviour. It can be called as deduction.
3. If our "human components" are that easy to put into logical elements, how are we different form machines?
"At maturity, the estimated average number of cells in the human body is given as 37.2 trillion." Function of most of them we do not know. How you can imagine that it will be fusible to emulate only a part of the human. Humanoid will always remain substitute of the human. In my opinion it make no sense to replicate human. We have to focus on building thinks which can improve our live not replace it.
You are saying that part of intuition behaviour is deductive thinking - I would say not deductive (logical process, based on statements like "if A, then B", leading to one, certain conclusion) but abductive (also a logical, based on testing hypotheses / inference, but not leading to one certain conclusion). I think we will have a chance to discuss this topic together this friday :-)
DeleteI would put hypothesis that intuition is far far older and not only monopolized by human species.
Delete1. Are human emotions really that simple that they can be put apart into logical elements and easily learned by a machine?
ReplyDeleteHuman emotions are very complex. Machine can distinguish some basic behaviours and expressed emotions (six types, according to comment above) but in my opinion it's impossible to learn all them without having a soul or advanced artificial intelligence.
2. Can we do the same with our intuition? What could the design of intuition look like?
Intuition is something acquired over a years of evolution with addition of experience gained during human life. It could be possible to design how it looks like after deep study in that field.
3. If our "human components" are that easy to put into logical elements, how are we different form machines?
As I already mentioned in answer to the first question - we heave souls :)
In my opinion it is not easy to simplify human emotions in order to remodel them. These are very complex compositions and we cannot do the same with intuition (I hope so).
ReplyDeleteHowever, it is only my opinion. I really hope that this part of humanity will not be dissected.
We differ from machines because we often behave illogical and we are not obvious.
1. Are human emotions really that simple that they can be put apart into logical elements and easily learned by a machine?
ReplyDeleteI think that basic human emotions could be learned by a machine, but it is not often happening that we feel just one simple emotion. People are struggling with many emotions and "mixed feelings" that are difficult to distinguish to other people, not speaking about machine.
We are far from the situation when computers will understand our emotion, or it may even never happen.
2. Can we do the same with our intuition? What could the design of intuition look like?
If we define intuition as some kind of internal prediction mechanism they it could be implemented as some algorithm ar set of algorithms. Even now there are tools to predict future situation on an basis of historical data and actual values of some variables.
3. If our "human components" are that easy to put into logical elements, how are we different form machines?
I disagree with the first part of this question. "Human components" are not easy to put into logical elements. Scientists are working for decades on creating structures similar to human brain and they are still far from this.
I don't think that human emotions are that simple. Machines can pretend that they can understand them and express their own emotions but as for now it's mostly based on tricks. From my point of view, machines differ from humans fundamentally. Machines can emulate emotions but their way of operating is purely logical. In case of people emotions prevail.
ReplyDeleteThis article is amazing. I am interested in modern technology but I am not interested in the part about artificial intelligence. I have ever thought that real robots are simply machines with short lists of things to do. After reading this article I realized that robots can do all the movements which are included in their entire software. But the issue concerning the problem which robots can do such things is quite different. Earlier it was impossible for me to realise that robots can recognise our emotions! It turned out that this is not impossible to teach robots to do that. Fantastic and strange. According to SF films in the future robots will be able to take our position in Earth. But I think it is very difficult to teach robots the appropriate recognition of human emotions. I think this is too difficult to programme machines to recognise all people emotions. In my opinion intuition is too difficult for researchers to be recognized. I think intuition like our soul is impossible to copy because this issue is too difficult to be described in maths symbols. In that way I think our humanity is safe. So far.
ReplyDeleteIf robots are to overtake our position in the world, then they will not need to recognize our emotions correctly, right? :-)
DeleteI don't agree with you - master must recognize his slave emotions.... ;-)
DeleteThis comment has been removed by the author.
ReplyDelete1. Are human emotions really that simple that they can be put apart into logical elements and easily learned by a machine?
ReplyDeleteNo they aren't. Human emotions are very complicated and nobody knows how "it works" :)
2. Can we do the same with our intuition? What could the design of intuition look like?
No we can't. Maybe by sum of expirence but it will be not working.
3. If our "human components" are that easy to put into logical elements, how are we different form machines?
It not will be easy.
1. Are human emotions really that simple that they can be put apart into logical elements and easily learned by a machine?
ReplyDeleteUnfortunately the answer is yes. There is finite set of our reactions for each person determined with culture of studied subject. So each person can act predictably depending of environment and course where he grows.
2. Can we do the same with our intuition? What could the design of intuition look like?
With intuition is something different. I would set a hypothesis that this is more advanced mechanism constructed on learned patterns and deeper intelligence developed on experience of resolved assignments. In my opinion intuition is something like fingerprint of our brain with can be used to identify people which are known to us.
3. If our "human components" are that easy to put into logical elements, how are we different form machines?
I can't wholly understand when machines would be smarter than us, but still we humans are step forward than machines. We don't want to fall into machine optimization or frankly speaking eradication of human kind. So in this race we will develop human intelligence to maintain superiority over machines.
1. Are human emotions really that simple that they can be put apart into logical elements and easily learned by a machine?
ReplyDeleteHuman emotions are complicated. Each person express emotion in specific manner. The way each individual person experiences feelings is reflected in a method of expressing emotions. Scientists can put into logical parts only behavioral expression of emotion, in that way they are able to simulate emotional life. Machine is unable to feel, the only thing that it could do is to learn how to simulate emotions.
2. Can we do the same with our intuition? What could the design of intuition look like?
That's really hard question. The answer depends on understanding of term intuition. The design of intuition could be the generator of random numbers filtered by probability of future events. Source of that knowledge could come from Internet. That's why robot with intuition simulator should be in that situation always online.
3. If our "human components" are that easy to put into logical elements, how are we different form machines?
I cannot agree with that claim. In my opinion scientist can not put apart human components into logical elements, they can only simulate that using specially designed algorithms. I mean, it's self-awareness that separates humans from robots with good emotional expression ability.