Talk:Affective computing

From Wikipedia, the free encyclopedia
Jump to: navigation, search
WikiProject Robotics (Rated Start-class, Mid-importance)
WikiProject icon Affective computing is within the scope of WikiProject Robotics, which aims to build a comprehensive and detailed guide to Robotics on Wikipedia. If you would like to participate, you can choose to edit this article, or visit the project page (Talk), where you can join the project and see a list of open tasks.
Start-Class article Start  This article has been rated as Start-Class on the project's quality scale.
 Mid  This article has been rated as Mid-importance on the project's importance scale.
 
Note icon
This article has been marked as needing immediate attention.
WikiProject Computer science (Rated C-class, Low-importance)
WikiProject icon This article is within the scope of WikiProject Computer science, a collaborative effort to improve the coverage of Computer science related articles on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.
C-Class article C  This article has been rated as C-Class on the project's quality scale.
 Low  This article has been rated as Low-importance on the project's importance scale.
 
WikiProject Human Computer Interaction  
WikiProject icon This article is within the scope of WikiProject Human Computer Interaction, a collaborative effort to improve the coverage of Human Computer Interaction on Wikipedia. If you would like to participate, please visit the project page, where you can join the discussion and see a list of open tasks.
 ???  This article has not yet received a rating on the project's quality scale.
 ???  This article has not yet received a rating on the project's importance scale.
 

I think[edit]

I think it should also mention Marvin Minsky's book The Emotion Machine 2006

Interesting! Can we have some more on when work started in the field (Parry springs to mind) and what advances have been made recently ? -- Derek Ross 23:43 25 May 2003 (UTC)

Organization[edit]

I have prepared a project on this topic a few months ago, so I did a bit of research on the current papers. I have a lot of information of this topic, especially on recognizing affect and I will share the best of it, as the article lacks detailed information. What I could offer best fits in the technologies of affective computing section. I have already placed some info there, in the Physiological monitoring and Speech affect sections, and I have some more. However, a lot of sub-topics have been formed and I don't like the layout. Do you have any suggestions? Could you organize it a bit? Thanks S33us00n (talk) 09:00, 4 September 2011 (UTC)


Quality of article[edit]

This article is barely readable. Although the content is interesting and largely irrefutable it is terribly written. May I suggest a rehaul of the grammatical mistakes?

I've started, but I don't understand enough of the field to get right through it. It looks as though it has been translated from Japanese or something. Myopic Bookworm 11:15, 2 October 2006 (UTC)
PS The article needs to be prevented from heading off into more general issues of emotional research.

Agreed the quality of expression leaves something to be desired, but it does have a distinct advantage over a lot of other articles I read on Wiki: right near the start it gives a link to a full PDF of a seminal academic paper on the subject. And the link works! That the paper is by the leading author in the area can be independently confirmed, thus the article gave me precisely what I wanted - a short-cut to the conceptual basis of the subject, without having to wade through hundreds of derivative papers on Google Scholar. So the authors do deserve some credit. --Wally Tharg (talk) 10:00, 6 March 2010 (UTC)

Why Affect?[edit]

I have only recently stumbled upon the arena labelled as 'Affective Computing'. As a social scientist the term 'Affect' has a specific set of meanings denoted by a particular philosophical literature. It seems to me that what is presented in this article presupposes a conflation of 'affect' with 'emotion', which is rather contentious. What I have included below is some of my own grappling with the idea of 'affect', edited from a Masters degree dissertation [Kinsley S (2006) 'An End of Cyberspace? Metaphor, Affect & Socio-technical relations' unpublished: University of Bristol] in the hope that it offers some sort of inroad into reflecting on how or why one might use affect. I would certainly encourage all those taking 'Affective Computing' forward to consider the literature I touch upon. Furthermore, given these points, should those interested in such an arena of research consider alternative titles such as: 'Emotive Computing' or 'Preceptive Computing'? -- Sam Kinsley 6th Novembe 2006 12:05 GMT


A gamut of understandings of affect can be traced across various disciplines, from social psychology and studies of cognition to continental philosophy (for a review see Thrift, 2004). The implicit danger to be recognised in any discussion of affect is a confusion or simplification to ideas of emotion, a criticism that might be levelled here. Prominent amongst theorisations of affect is the work of philosopher Gilles Deleuze who develops the ideas of Baruch de Spinoza (see: Deleuze, 1988).

To productively follow this line of conceptualisation, we should commence from an understanding of experience as trans-subjective, in which, according to Deleuze & Guattari (1988: 180), ‘there is no longer a self’ and bodies vacillate like ‘a glowing fog [...] that has affects and experiences ovements and speeds’ (Ibid.). We can therefore understand affect as a pre-cognitive and impersonal capacity. Experience, according to (Deleuze & Guattari, 1994), can be understood in terms of ‘affects’, the pre-cognitive and perhaps involuntary recoil from a strong odour or the leap of the heart when faced with an attractive man or woman, and ‘percepts’, the odour itself or the appearance of the body. Affect, Thrift (2004: 60) suggests, can thus be understood ‘...as a form of thinking, often indirect and non-reflective, it is true, but thinking all the same’. Further:

‘...if you define bodies and thoughts as capacities for affecting and being affected, many things change. You will define an animal, or a human being not by its form, its organs, and its functions, and not as a subject either; you will define it by the affects of which it is capable’ (Deleuze, 1988: 124).

Do we need ‘affect’ to conceptualise our experience of the world? Indubitably, given that, to paraphrase Hamlet, there are more things in this world than can be understood, represented or conceived of consciously. Our relations are often motivated pre-cognitively and affect, therefore, provides a means of theoretically articulating such motivations.

Some references:

Deleuze G (1988) Spinoza: Practical Philosophy, Hurley R (trans.) San Francisco, CA: City Light Books

Deleuze G & Guattari F (1994) What is Philosophy? Tomlinson H & Burchell G (trans.) London: Verso

Deleuze G & Guattari F (1988) A Thousand Plateaus, Massumi B (trans.) 2004 edition, London: Continuum

Massumi B (2002) Parables for the virtual: Movement, affect, sensation, Durham, NC: Duke University Press

Thrift N (2004) “Intensities of feeling: Towards a spatial politics of affect”, Geografiska Annaler 86 B (1): pp. 57-78

Sam, this article conflates affect and emotion in a way that the academic study of affective computing does not. Affective computing is concerned with human response in the Deleuze & Guattari sense you mention above rather than with emotion as the article seems to indicate, but it is a common simplification for introductory lectures or popular articles on the subject. If you can improve this article to make the difference clear in a way that won't confuse those who are new to the concepts, that would be great! --Siobhan Hansa 13:28, 6 November 2006 (UTC)
Siobhan, thank you for your quick response! I am not sure I am the best person to edit this page. The above quick discussion is a rather hackneyed version of a section of my own Masters dissertation and certainly to brief to address the broader discussion that would be necessary of Affect given the application to computing. Whilst this is a fascinating concept, and one that is close to my research interest, I am still a little confused by calling the arena of study 'affective computing'. Firstly, it seems to me that there is a wider conflation of 'affect' and 'emotion' outside of this article, for example: 'Affective Computing is computing that relates to, arises from, or deliberately influences emotions.' MIT Media Lab Affective Computing Group. Beyond this, a Deleuzian reading of affect might sugest that you cannot empirically access affects in and of themselves, only the resultant perceptual cognition(s), examples such as this are why I suggested a revised title. I would be very interested to hear from an 'Affective Computing' researcher on this. -- Sam Kinsley 6th November 2006 17:24 GMT
I didn't mean to imply that there is a direct equivalence between the use of affect in comp sci and affect in social science. Just that - though emotion is frequently used as though it were interchangeable with affect, affective computing looks at response in a broader sense than the definition of emotion that you are using here. It may be worth remembering that many fields have slightly different understandings of similar concepts because the desired outcomes from study and the presures within the fields are different. That doesn't mean it's not worth considering what other fields have to offfer, but to go from that to a suggestion that one field should change its name because it doesn't adhere to another field's definition of a term is a little parochial. --Siobhan Hansa 18:04, 6 November 2006 (UTC)
All very good points. I'm sorry, I didn't mean to suggest I was espousing a peculiar parochial one-up-manship, rather I am simply asking: Why use the term 'Affect' in this case? As I suggested in my first section above there are many disciplinary takes on 'Affect', I am confused as to how 'Affective Computing' needs 'Affect' rather than 'emotion', which would be equally as worthy a subject of study - just different. I hope that clarifies my concerns - I certainly do not want to claim a peculiar disciplinary authority or preclude any perspective from an interesting and reasonable discussion :-) -- Sam Kinsley 7th November 2006 12:08 GMT
Perhaps I was being a little defensive. I disagree that affective computing is simply about emotion. While there is a lot emphasis placed on emotion as a describer of what people are trying to achieve, the reality is computers can't read emotions or hold them. They can read things like whether someone types faster, or squeezes a mouse harder, or uses more negative or more positive words. Computers can behave in one way rather than another - respond faster or slower to users, use different colors, words or graphics, or bring up different information. But beyond these types of indicators (and ways to manipulate them), affective computing also looks at things like motive (how do you understand what a user's motive is, or how do you give a computer motive). Although none of these things are simply about emotion (or perception), emotion is frequently used as a shorthand in this sort of situation.
I don't think I'm doing a good job of making a case for supporting the use of affect over emotion to you. Perhaps because I have less understanding of your own field and the language you use than I need. The Swiss Center for Affective Sciences describes what I consider to be affective computing's approach and places it within a broader academic constiuency. It might be useful to you in understanding how other branches of academia are conceptualizing this aspect of human behavior. --Siobhan Hansa 17:11, 10 November 2006 (UTC)

Dubious tag[edit]

I deleted this:

(see strong AI). The possession of innate emotion in non-human intellects is primarily a philosophical topic, since sapience is considered a pre-requisite for the ability to process emotions and there are currently no known models of sapiency besides humans.

Here are my objections: (1) why would strong AI be necessary to process the semantic differentials used to model affective meaning? I think you've got the argument backwards: the claim is that affective meaning is necessary for strong AI, not the other way around. (2) It uses the sci fi word "sapience", which is meaningless in this context. Do you mean "consciousness"? (which has a meaning for neurology or philosophy). Do you mean "intelligence"? Do you mean "soul"? ---- CharlesGillingham (talk) 10:35, 17 November 2007 (UTC)

Doubtful history[edit]

I am extremely doubtful that the field of affective computing started with Rosalind Picard. This sounds like an extreme exaggeration at best.--Filll (talk) 01:17, 6 May 2008 (UTC)

I might also note that it seems to me that these activities were going on in Japan, long before Picard was involved. And maybe at IBM and other places too. So, I have my doubts. Maybe she invented the name. Maybe. But the field? --Filll (talk) 02:39, 6 May 2008 (UTC)
I invented neurothanopathy. So? •Jim62sch•dissera! 02:43, 6 May 2008 (UTC)
Funny you FILL the article with "vague" tags and then support this with a vague assumption :) Did you read the PDF? Probably not because it would answer the question. Yes others might have (you didn't show any citations for your statement so I'll have to include the "might" here) done some research in the field before, however she formalized it. Thus the field of affective computing really did start with her. I hope you guys enjoyed mutilating this article. 195.216.82.210 (talk) 11:31, 6 May 2008 (UTC)
Sure. We'll just go by best knowledge, that works fine. Can you cite earlier sources? --Kim Bruning (talk) 13:51, 6 May 2008 (UTC)


I am not an expert in this field. But I do remember reading about this back into the 70s and even into the 60s. And I have a huge number of links I have been going through. If we are going to have an article on this, let's do the best job we can instead of just creating a disgusting inaccurate mess with no references and no sources.--Filll (talk) 01:20, 7 May 2008 (UTC)

Pppbbtthhhh[edit]

I removed the following because it presupposes much yet proves naught.

===Emotional understanding===
Emotional understanding refers to the ability of a device not only to detect emotional or affective information, but also to store, process, build and maintain an emotional model of the user. The goal is to understand contextual information about the user and its environment, and formulate an appropriate response. This is difficult because human emotions arise from complex external and internal contexts.
Possible features of a system which displays emotional understanding might be adaptive behavior, for example, avoiding interaction with a user it perceives to be angry. A probable use of such capability would be ensuring data integrity and security.

Sorry, but you've gotta prove this drivel. •Jim62sch•dissera! 01:27, 6 May 2008 (UTC)

<scratches head> What's the problem with it, according to you? --Kim Bruning (talk) 13:50, 6 May 2008 (UTC)
I already explained, I believe: "Sorry, but you've gotta prove this drivel". That is to say, id est, it is vague and uncited. Even the title is unclear: is it about understanding emotions or understanding something via the emotions? What precisely" does this mean and how is it meant to be accomplished? "A probable use of such capability would be ensuring data integrity and security." •Jim62sch•dissera! 17:37, 6 May 2008 (UTC)
I basically parse the paragraph as follows: The paragraph defines emotional understanding as a (future/projected) capability where a machine will be able to gather enough data to generate an internal model of the emotional state of a human being. It then provides an example where an adverse emotional state in a human operator might lead to loss of data integrity (and detecting such a state can therefore prevent such loss).
Oh, a cite? Ok, when I have some time to ask some folks to help :-) --Kim Bruning (talk) 20:26, 6 May 2008 (UTC)
Cite it when you can. emotional understanding is a future concept?
Ah, so we can detect the state of the human operator? What if the person is a con-man? •Jim62sch•dissera! 21:16, 6 May 2008 (UTC)
Apparently it can be worded more solidly still, and someone is looking up a reference for me.
My apologies for the poor wording. I don't think any computer is currently fully capable of "understanding" human emotions, though it might be possible to program them to do that in future, to some extent. Also, one can actually only detect the state of the computer model, obviously. If the operator is a con-man, the model will be flawed :-P
Anyway, I have no idea where you're headed with that.
In the mean time, I'll just ask someone to look up some sources for me, which are more adequately carefully worded. (And in fact, they're already doing so, I told them to take their time :-) )
--Kim Bruning (talk) 00:51, 7 May 2008 (UTC)
Hmm, strictly these folks seem to stick to Affect (the display of emotion in posture and facial expression), rather than emotion per-se, which is somewhat more easily measurable. But I might be sticking my foot in my mouth, I'm going to wait for proper sourcing, thanks. :-) --Kim Bruning (talk) 00:55, 7 May 2008 (UTC)
Given that not all humans can reliably detect emotions "properly", given that the expression of emotions varies widely, and given that humans program computers, I'd have to think that the odds of a program correctly discerning emotions correctly with absolute certainty, is, well, nil. But, feel free to find sources. •Jim62sch•dissera! 10:07, 7 May 2008 (UTC)
Given that the odds of anything happening with absolute certainty are nil, that would not be surprising at all. Now as to less than absolute certainty, that's where things get interesting. :-) --Kim Bruning (talk) 14:24, 7 May 2008 (UTC)

Tags[edit]

Excessive fact-tagging is considered disruptive. When I have time I'm going to go through this article and remove some of them. There's already a {{inline}} on top. --Relata refero (disp.) 11:19, 6 May 2008 (UTC)

Excessive use of unsubstantiated claims is non-academic. It is important to show precisely what claims need to be supported. That's the way of Academe, and as Wikipedia has, through the Foundation, expressed a desire to be considered an academic medium we must follow the same strictures. •Jim62sch•dissera! 17:31, 6 May 2008 (UTC)
There were so many statements that lacked substantiation that it had to be clear how bad it was. OrangeMarlin Talk• Contributions 17:43, 6 May 2008 (UTC)
No. That is actually exactly what article-wide tagging is for. --Relata refero (disp.) 08:49, 7 May 2008 (UTC)
You's think that was obvious, but, apparently not. •Jim62sch•dissera! 21:17, 6 May 2008 (UTC)
Meh, see above for why not. --Relata refero (disp.) 08:49, 7 May 2008 (UTC)
uh, where? I like specifity.  :) •Jim62sch•dissera! 10:00, 7 May 2008 (UTC)
Article-wide tagging is generally used for articles that are short of inline citations. See the recent TfD of nofooters for general custom on this. --Relata refero (disp.) 10:35, 7 May 2008 (UTC)

Autism, AC and Prostheses"???[edit]

Huh? Prostheses for Autism? "Affective computing is also being applied to the development of prosthetic devices for use in alleviating autism.". Obviously this is meaningless without further explanation. •Jim62sch•dissera! 17:39, 6 May 2008 (UTC)

Hmmm. I know an autism specialist who has a prosthetic arm. Does that count? Guettarda (talk) 17:59, 6 May 2008 (UTC)
I've read a few things about this work, it's really good, I'd dedicate a section to it. I would suggest that aside from the rather heady philosophical/theoretical world of AI, computers are applied as a communications tool (related to writing), and hence affective computing's chief application should therefore be as a communications device, either enhancing able peoples' capacities to communicate, or to enable people with impairment or disability. Here's a good starting place. http://affect.media.mit.edu/projectpages/esp/ ;-) you see what I mean? :-| ... :-) ––– --welcome to the 21st century (talk) 10:25, 11 February 2010 (UTC) bishopdante

Hmmm, [1] (I presume new tools would be "prosthetic" in nature), um, and [2] from this publication list, several of the 2008 papers look like those describe prosthetics. Does anyone have a science library in the neighborhood to check those? --Kim Bruning (talk) 20:31, 6 May 2008 (UTC)

We might consider them to be orthotic rather than prosthetic if they are intended to assist or correct brain function rather than replace the brain. (Curiously, I used to have the same argument with Doug Lenat about Cyc.) Bovlb (talk) 20:54, 6 May 2008 (UTC)

Hmm, one of my daughters is autistic (thus I have a dog in this chase); however, the term prosthetic is mierda del toro. •Jim62sch•dissera! 21:20, 6 May 2008 (UTC)
So orthotic is (more) correct? --Kim Bruning (talk) 21:43, 6 May 2008 (UTC)


Uh, are we talking about the same disorder? Orthotic is as incorrect as prosthetic. •Jim62sch•dissera! 09:58, 7 May 2008 (UTC)

Second para of lead[edit]

The second para of the lead says

Affective Computing is also the title of a textbook on the subject, by Professor Rosalind Picard, published in 1997 by MIT Press.[1] The origins of the field trace back to Picard's 1995 paper on Affective Computing

Do we really need that in the lead? Aside from the honorifics and the publication data on the book, isn't this more of a hatnote? Guettarda (talk) 18:07, 6 May 2008 (UTC)

I added "maybe". •Jim62sch•dissera! 21:21, 6 May 2008 (UTC)
I've added two sources which name Picard as the originator of this field, and removed the "maybe". I'm also working on collecting references for other {{fact}} requests, though this may take some time. In my casual reading on this topic, there seems to be nothing blatantly wrong (several references to both autism and "toys" in the literature, for example), so please be patient. ATren (talk) 14:30, 13 May 2008 (UTC)
The sources I glanced mainly cite her 1997 book as the main contribution, not the 1995 tech report? What is your impression? Merzul (talk) 19:14, 13 May 2008 (UTC)
I don't know - the two sources I cited seemed to be in conflict: Wired said it was the book, but the other (academic) source said it was the tech paper. They're both named "Affective Computing" so perhaps the book was an expansion of the paper? The paper came earlier, and it is cited by others, so perhaps that should be stated as the true origin. But it seems like a minor issue to me, since both are Picard's works anyway. ATren (talk) 20:57, 13 May 2008 (UTC)
Maybe just say it originated with the work of Picard? I don't know, and you seem to know much more than me. Merzul (talk) 21:29, 13 May 2008 (UTC)
Actually, I just started researching it yesterday, so I only have a day's worth of expertise over you. :-) I think it's fine to say it originated with the work of Picard, and if I find something more specific and definitive, I'll change it. I plan to work on this article in the coming weeks, to resolve the fact tags. ATren (talk) 21:58, 13 May 2008 (UTC)

"emotions in machines might be associated with abstract states associated with progress (or lack of progress) in autonomous learning systems." And it might not. This is really, really speculative. It kind of presupposes that we know the solution to how autonomous learning systems might "think", and might even risk anthropomorphism. I think the article would be fine without this sentence. Themusicgod1 (talk) 12:39, 15 July 2008 (UTC)

{{subst:aviso FP|Computación afectiva}}