Can robots in any case learn and adapt human emotions?

Question

Robots are just an advance programmed devices which does the task which are ordered to them via codes. Is it even possible for a highly advance robot to interpret and learn human emotions?

Update: From the suggestions from various comments-here is the more elaborated perspective of my world

Time: Any time in the future
Assuming only Human race exists (and no aliens) which has advanced technologically very well and are efficiently developing robots which can predict, sense and even visualize the human's brain, lies and his probable future actions.

Scenario: In due course, the robots which are programmed to predict and sense human's brain-subsequently starts interpreting human's state of brain (which is called "mood", "feeling", "emotion"). 'Learning' in the question refers to 'interpreting' and 'accommodating within' the same.

Since this is world building, answer to my question could be yes-but I want my world to be realistic and synced with present human race, technologies and future we are predicting.

Answer 1

Hard Science Answer With Current Technology

As a researcher who's familiar with some of this work (my last project was in brain computer interfaces and a colleague coded a system that used computer vision to evaluate human emotions), I am going to go ahead and give you a hard science answer based on current research even though you didn't ask for one.

Can robots sense human emotion? Yes, they already can (to a limited degree)

First, its helpful to realize that humans actively broadcast (communicate) many internal emotional states in ways that can be seen and heard - and recognized by computers. Many other emotional states can also be partially discerned by other physiological measures which are less obvious, but can be recorded by non-invasive means (IE, no need to install a neuro-jack into someone's skull or implant chips).

Vision-Only Emotions

There are a number of advances in understanding how humans recognize emotional states, such as a test called Reading the Mind Through the Eyes, which shows that humans can detect a wide variety of emotions (36 distinct emotions in one analysis, ranging from "playful", "skeptical", and "hostile") based only on looking at pictures of the eyes and near-eye areas. In computer vision there are projects to detect and convey human emotions in computer animations using only facial expressions, and this similarly has been shown to successfully recognize many emotions - and to be able to re-broadcast them in a way that humans recognize.

Computers can already recognize many human emotional states through simple, cheap web cameras and live computation. Accuracy is not presently perfect ( either in humans or in computers), but if we already have this now than naturally advances can only improve the range of detected emotions and the accuracy of detection. Currently most technology uses only a single freeze-frame (or just a few frames) to make this detection.

Audio-Only Emotions

Without even looking at the meaning of words, there are computerized methods of real-time detection of human emotion using only audio voice recordings. Just one example of this is EmoVoice (check out their project for more information) - and this included emotionally appropriate animated facial response to human speech...in 2005.

Again, more work can only advance the accuracy and range of emotions that can be detected.

Beyond Human Limitations - Brain Activity, Heart Rate, Skin Conductance, etc

The above methods are already currently useful and have full potential for going beyond human levels of accuracy in some limited-context scenarios - but what if we used information that normally was not available to humans?

Using wearable brain computer interfaces, such as EEG headsets (which detect electrical activity from the brain on the surface of the skin), we can already use a computer to identify a variety of emotional-cognitive states, such as attentiveness (is a driver paying attention?), sleepiness, positive/negative emotional reactions, etc. Having worked on areas related to this I can assure you that this is, in fact, hard to do reliably with current technology, yet there are many methods that have been shown to actually work! Emotional recognition is hard, but many advances are being made.

In one talk I attended just this year, Conceptual Priming for In-game BCI Training, an experiment showed that using only EEG a computer can even detect mentally visualized images - such as distinguishing between whether a person is thinking about a flashlight or a gun or neither, with about 60%+ accuracy based only on a single implicit training presentation lasting a few seconds. This is not an emotion, but if a wider range of images can be recognized then this could also be used to infer emotional states - or just obviate the need to detect a lie at all.

Theoretically you could ask someone "where did you hide the body?" and they would implicitly think of where they hid the body - and you could capture that snapshot directly from their brain. We are a long way off this being practical in the next 10+ years, but given 100-200+ years then this is extremely likely to be possible to do based on our current science - especially if methods similar to fMRI can be expanded to better detect human visual activation in the brain.

Other methods of emotion detection include things like examining skin conductance and heart rate. Alone these methods have proven more limited in the ability to detect emotion - but it has been shown to have connections to detecting stress levels, excitement, emotional valence, and emotional responsiveness. This information is emotional in nature, and some current research looks at the viability of combining these measures to detect a wider range of emotional states with greater accuracy - but the results aren't really "in" on this yet.

fMRI has also shown promise in detecting and evaluating emotions, but the technology is less practical currently as the machines are large and expensive and require a great deal of experimental control. In the future if this technology can be shrunk and made more resilient and easier to use, it may be a new field of computerized emotional state detection.

Put It All Together

I am not presently aware of research that has successfully combined multiple modes of emotion detection together - such as face expression, audio, brain activity, skin conductance, heart rate, etc. It would extremely likely be able to greatly improve the ability to detect a far larger range of human emotional states and responses to stimuli, and given enough computational power could do this all live within a single robot in the not-distant future.

Finally, I will leave you with one more set of keywords. There is a presently emerging field specifically dedicated to using computers to detect, respond to, and even "have" emotions: affective computing. To quote from the MIT Media Lab's Affective Computing Group:

Affective Computing is computing that relates to, arises from, or deliberately influences emotion or other affective phenomena (Picard, MIT Press 1997).

Emotion is fundamental to human experience, influencing cognition, perception, and everyday tasks such as learning, communication, and even rational decision-making. However, technologists have largely ignored emotion and created an often frustrating experience for people, in part because affect has been misunderstood and hard to measure. Our research develops new technologies and theories that advance basic understanding of affect and its role in human experience. We aim to restore a proper balance between emotion and cognition in the design of technologies for addressing human needs.

Our research has contributed to: (1) Designing new ways for people to communicate affective-cognitive states, especially through creation of novel wearable sensors and new machine learning algorithms that jointly analyze multimodal channels of information; (2) Creating new techniques to assess frustration, stress, and mood indirectly, through natural interaction and conversation; (3) Showing how computers can be more emotionally intelligent, especially responding to a person's frustration in a way that reduces negative feelings; (4) Inventing personal technologies for improving self-awareness of affective state and its selective communication to others; (5) Increasing understanding of how affect influences personal health; and (6) Pioneering studies examining ethical issues in affective computing.

Affective Computing research combines engineering and computer science with psychology, cognitive science, neuroscience, sociology, education, psychophysiology, value-centered design, ethics, and more. We bring together individuals with a diversity of technical, artistic, and human abilities in a collaborative spirit to push the boundaries of what can be achieved to improve human affective experience with technology.

---

TLDR; Current technology already allows a wide range of human emotional detection by computers/robots using many available measures, from the human-like (facial expressions, sound of voice) to the beyond-human (brain activity, etc.). Additional research can only extend the range of emotional detection, improve accuracy, and expand our understanding of human emotions. There is no question that robots can possess various degrees of this ability in the future, because they already can right now.

Answer 2

Asking whether robots can learn emotions is the kind of question that will very quickly become philosophical. But just answering it at face value:

Yes, they could.

Human emotions, for the most part, are based on our state of mind, our sensory perception, and a number of goals that we consider important. Computers could be given the same goals, and their responses would likely be impossible to tell apart from our own; they would behave with emotions. Whether someone calls those emotions "real" is, like I said, philosophical.

The reason that computers currently don't have emotions is mostly because all the goals that humans have that cause the emotions are ancient, and based on evolution. Evolution has taught anything that lives that important goals include "propagate" and "stay alive/safe". Computers don't have these.

(And, for somewhat obvious reasons, most humans would not enjoy the idea of giving robots these goals)

The end result is that computers don't have these emotions naturally and we'd need to teach them, but even we don't really understand emotions enough to do it. So that makes it very hard for us to artificially give emotions to computers, and our own emotions makes us unwilling to try and give them natural emotions.

(There's a good chance that if you build a set of computers that can self-propagate, self-modify and you give them a long period of time to do it in, that the resulting designs will have emotions. There is also a good chance that this uncontrollable machinery will be dangerous.)

Answer 3

A "mood" is simply a specific state of mind that alters the rules by which we act and react to stimuli.

By that definition, robots already have "moods". This becomes especially apparent with for instance robotic spacecraft. When something goes wrong with them, they go into a sort of "panic mood", although it is not called that, it is called "safe mode". In this state, the spacecraft does pretty much what a human does in a state of panic: focusing on the bare essentials of staying alive. Anything that is not considered essential for staying alive is ignored for the moment.

So can it happen? Yes, it can. If the simple "dumb" robots of today have the evolutionary seed of "moods", it follows that their more advanced descendants will probably also have them.

There is absolutely nothing to suggest that once robots reach the state of intelligence that we could consider them equal to us, that they also cannot have moods, or learn to interpret them in other intelligent entities.

Answer 4

Yes and no.

This gets down to philosophy summed up in "Is your blue the same as my blue?" The way we answer that is that we define blue as photon wave frequency so that blue is a photon at the energy level of some 100s of hertz.

With emotions what you have to realize is that they are generated by chemicals being released and received based on some neurons firing and when those chemicals are received some other neurons are activated. We can certainly duplicate that. It's just systems interacting and sending signals to each other.

The problem is what triggers those chemicals and what does "x" feel like. What does "pleasure" feel like? I don't know, but it's pleasurable! We can program, for example, you've done "x" there for you get a +1 to the "pleased" variable, but we don't have any way to quantify how "pleased" feels within each other, so even if we replicate the human brain electronically perfectly, we still can't say whether they are feeling anything like what we're feeling, and by we, I mean me, because I don't even know what you feel like when you're feeling good. Your "Feels good" could literally be my "Feels very painful" and in some people, it is.

Once you go to the even more abstract of Happy, sad, etc. you can give someone a list of things that you mean when you say happy and they can say yay or nay, but reality is that what they do or feel, that defined thing is the complete opposite of what you're feeling.

The only thing we can really say is that we can make it act like it has emotions to ours and we can only say our emotions seem to correspond to similar things most of the time using this similar way of operating so they probably feel within the same range of emotions as us, but we can't ever know for sure.

Answer 5

Anything is possible if it exists in nature.

That we are (currently) unable to replicate it is irrelevant, because we could advance to a level of understanding at which we are able to.

Evolution

Possibly it's of a complexity that lies beyond human comprehension, in that case we could implement code that is able to learn and mutate itself with a specific goal in mind. Building up slowly, gaining complexity as it grows.

Emotion is a big inherent factor in how living beings work, it's a crucial part of each being's inner fabric. So, if evolution can invent it, we should be able to as well and likely far faster, as we can design mutations with a specific goal in mind.

Possibly at some point we could read 'human coding' and copy the same 'emotion routines'. Many inventions have been created by studying what already exists in nature.

Is it easy? No. Is it possible? Very much so!

Answer 6

So, what is a mood? Ignore all the warm feelings you get from being in love, or the fire you feel when you absolutely hate someone - Ask yourself what these emotions do, and what triggers them.

Anger/Hatred

• Triggered by a perceived wrong or hostility towards an individual, or by seeing behaviors deemed negative and harmful to the units community
• Results in less charitable actions towards its target, up to and including direct physical violence.

If something is harmful to you, it makes sense to harm it right back. If all humans did was run away from lions instead of turn around and spear them, we'd be no better off than impala. Anger seems to be a kind of active self-defense. After all, people can't harm you if you don't give them a chance.

Fear/Panic

• Triggered by an immediate existential threat
• Results in actions geared towards self-preservation

As mentioned by Michael Karnerfors, Spacecraft already do this. To an extent, computers do too - a blue-screen or kernel-panic is just a computer sensing out-of-tolerance or dangerous conditions and taking measures to preserve its hardware.

Happiness/Joy

• Triggered by some form of success, positive change, positive environment, or helpful actions from another unit.
• Results in more sociability/actions beneficial to the units immediate community, and taking 'celebratory' actions that might otherwise be considered wasteful.

Ever notice how when someone's happy they're a bit more cavalier with their spending? They might take their friends out to the movies, or at least treat themselves to something. The reasoning seems to be "Things are good, so I can afford a little bit of excess." That, or they want to spread the well-being to their friends.

So, with all of that, would it be possible to make a robot have Emotions? Yeah, probably. What we do with neurochemicals they could do with state-variables. Humans are already a pretty good model for this because we're individual, autonomous units that function best in a group.

I'm not sure an AI could learn emotions, but given enough time a group of AI that need to co-exist might develop something similar.

Answer 7

A difficult question. I am not so sure if it is possible. Of course it should be no problem to implement a kind of emotional response (if someone is yelling at you be sad; or if someone is petting you be happy). But the question is, are this actual emotions or is it just an imitation of emotions. If I tell you that you have to be sad if it is raining, you don't have to be sad eventually. Perhaps once we have developed an AI that is self-adapting and evolving, the AI may become emotional; it may emerges emotions and awareness of the self and finally pursue self-preservation. But a "simple" robot that has only the purpose to serve the humans without being aware of itself and its wishes, dreams and demands will not have actual emotions.

Answer 8

What makes an emotion a human emotion?

It can only be experienced by a human.

How are humans different from other living things (presumably)?

Humans are conscious*.

What is consciousness?

In his Magnum Opus Godel, Escher, Bach: The Eternal Golden Braid Hofstader argues that consciousness is an emergent property of the system of neurons in one's brain. This simple (at the micro-level) system of electrical potential mediators has the potential for complex phenomena in no way implied by the components of that system themselves. There other natural incidences of this phenomenon of emergence listed in the book.

In his book on artificial intelligence, Jeff Hawkins (inventor of the Palm Pilot) argues that building a brain analog is the only way we will create artificial 'human' intelligence.

So to answer your question, if the humanity of a human is indeed an emergent property of the nervous system, then yes absolutely computers are capable of being 'human' even if our present level of technology is incapable of building such an architecture.

*You certainly can argue my definition of 'humanity' here and throughout but for the purposes of this question I believe it suffices.

Answer 9

They can (as was covered in other answers), but most likely not via the brainwave-reading technology that you imply by "sense and visualize the human's brain":
- There are two different types of synapses, electrical and chemical. The latter ones do not necessarily induce electrical responses, so their activity is nearly impossible to detect from outside the brain.
- Emotions are especially difficult to detect, because they are hugely depended on the hormonal (=chemical) state of a person.
- You'd need a presize tracking of the source of every wave, as you'll get a lot of noise, both from brain and spinal cord - no only "thoughts" and "memories", but also the signals related to breathing, digestive system, standing/walking, etc.
- The skull damps and blurs brain's electromagnetic signals. Tinfoil hat worsens the situation.

So, if I were you, I'd simply make the robots recognize humans' emotion by old good facial expression, voice modulation, body language, and things the person says/ does not say.

Answer 10

The answer is very clear, and resoundingly so:

No!

Why not?

The reason is very simple: humans are not machines. A robot is a machine.

The most general definition of "machine" is a form of automatic apparatus which has been programmed in one form or another to react to certain stimuli. But no matter how complex, no matter how fine-grained this programming may be, it is yet programmed to be automatic, and has no power of choice.

A human, on the contrary, has power of choice. He is not a machine. He is ALIVE. Or, we may have to elaborate on this - he consists of a "thing", the body, which actually IS a machine, a biologically built machine. And then there is his mind which is the essence of LIFE - and it is capable of CREATING automaticities and machines. (It can also DECIDE to be very automatic which is, probably, why some people think that man could be an automaton ...)

However, a mind is by itself no machine. Have you ever dealt with children? Are any 2 of them ever alike? Can you ever predict what is going to be the next thing they are going to do, in their whimiscal ways? No, you cannot. (Unless you are only around psychotics - and yes, they are predictable machines, as the medication and electroconvulsive shock therapies have reduced them to a status of automaton.)

What do you think - would a robot be capable of experiencing aesthetic pleasure created by listening to wonderful music? Would a robot have been able to create Mozart's symphonies? Never! Creativity is a LIFE thing.

The "brain" is NOT the mind of man. This is a typical misconception of current western materialism. The mind uses the brain to get things done by the body - but I am telling you, you are NOT thinking with your brain. Maybe when you do boring chores like driving a car - the mind has delegated that chore to some circuit. But when you CREATE, when you deal in AESTHETICS, you do not use your brain. Have you ever really created anything? People who think man is a robot must have a very poor outlook on life, because if you actively create you KNOW that you are not a machine.

So all people with free minds - be happy, because you ARE NOT A MACHINE!

Answer 11

Yes, the same as humans. Sometimes even better.

I believe that computers can do any thinking process that humans can - the only thing that limits them is their architecture("source code", if you will), and processing power. Both are irrelevant if we are in distant future(or just SF setting).

This has already been done, partially - robots can (at least partially) recognize emotions from voice or pictures/videos. What you're asking is a bit more complicated, but we are not that far away from this. I am pretty certain that within 50 years we will have at least primitive robots/programs that do this - and by primitive, I mean as primitive as current state of the art dialog agents (Google Now, Svoice, Siri, ...).

We have currently programs that can detect or predict some things. We've got Recurent Deep Neural Networks(check out for example this link). They aren't that good, but look at how new these things are. In 300 years these will be ancient technology - we're bound to invent something much better.

As for detecting lies, predicting the behaviour and "magically" detecting mood - depends. Depending lies and predicting the behaviour needs some samples - you can't predict it without some external knowledge(e.g. habits of the person). Then it could be trivial(some people might do something each time they lie), or impossible, depending on the person, but doable. Detecting mood or lies by just listening to the person - if it's not something that human could detect - I'm afraid there's slim chance robot could.

Answer 12

Answer to original question: yes programs have been able to reliably detect or interpret human emotions, usually based on a picture of their face. You can easily find tons of videos of this. This works by looking at features such as raised eyebrow, shape of mouth, ... That sounds rash but can be made very precise.

Now you ask if programs/robots could learn this by temselves? This is what is called unsupervised learning. For this to work, there must be a feedback mechanism so that the program can have a clue whether it guessed right or wrong. In real life a human usually has positive or negative feedback when she righly or wrongly guesses the anger, sadness, joy state of another human being. Without such a feedback it is mostly impossible for a program to learn properly. But with such a feedback, there is no reason it cannot learn. That would probably imply the robot has an engaging body, and/or masters spoken language.

Answer 13

There are several plausible scenarios to achieve this. One extreme example would be...

1. Create a software simulation capable of mimicking biological evolution (similar physical laws and environment to what we find on our world).
2. Observe and selectively control the evolutionary process so that it matches what we believe happened in our world, but selectively steer the evolution toward your desired machine (force the evolution via silicon, or some other molecular structure). Where the outcome doesn't satisfy, make adjustments and re-run the model.
3. Continue tweaking the model until your desired creatures have fully evolved with the characteristics you desire (emotions, empathy of their own so they won't immediately kill us all, physical and emotional compatibility with this world, interaction with humans, cybernetic upgrades, etc.).

Steps 1-3 should happen inside a quantum computer – shortening the process from billions of years to a few days.

4. 3D print the mathematical model onto a physical, synthetic representation (consider printing it from the embryonic stage and letting it grow)

You now have a machine that feels, but at that point it won't match the modern definition of a "robot." Also, the simulated beings will very likely have subjective experiences of joy and suffering within the simulation that are no different than our own. Evolution by natural selection is not pretty.

Could your world use this process to make a robot that feels? Yes.

Should it do this? No.

There are more ethical options that do not simulate evolution but do still involve the process of mathematical modeling -> simulating/testing -> refinement. The execution environment where the "robot" thinks (the software program runs) could either reside in a prefabricated neural structure or the entire neural structure could be 3D printed as a part of the program. The latter may be better conducive to the level consciousness you're seeking.

Answer 14

Given unlimited processing power, your robots could simply model the brain or whole body of the most talented negotiator or psychologist they can get their hands on. In fact, with unlimited processing power, they could simulate every atom in a room full of behavioral scientists, negotiators, and psychologists analyzing data in whatever division of real time you choose. Of course, it would probably be easier and less intensive to find some other way to model emotions, but what is the fun of that? I can just picture every robot having a room full of experts in his processing center squabbling about what exactly that minute muscle movement meant!