Area 51 » Science

264

followers

81 Example Questions (4 closed)

active newest votes

up vote 19 down vote

added by Quill Jan 14 '16 at 14:55

up vote 18 down vote

added by James Jan 7 '16 at 22:45

up vote 18 down vote

added by justin Jan 13 '16 at 10:04

On any SE site, this question would almost certainly be closed as "too broad". You are trying to answer the most complicated and overwhelming challenge in AI in the space of one post. That would not make a good question for Stack Exchange. – Robert Cartaino Apr 8 '16 at 15:20

No, this can be reasonably answered. Researchers had already programmed in the ability of Mario to simulate emotions, and the approach it uses is pretty similar to "The Sims" video game. AI can also identify emotions in other people too, using facial recognition technology. So it's not entirely impossible to answer this question, though you have to play fast and loose with the definition of "emotional intelligence". – Tariq Ali Apr 8 '16 at 23:10

I agree the question is "broad", but more like because it is open-ended: One answer today is probably not definitive as technology evolves. Closing without discussing is going too far imo (this is the most up-voted question...), but I do agree there is refinement work to do here. Thank you, Robert and Tariq. I for one would appreciate suggestions to break the question down, instead of just discarding. Perhaps @justin can rephrase based on Tariq's points? – Eric Platon Apr 8 '16 at 23:28

If @justin doesn't want to rephrase the question I could just submit this as two questions: (a) How do you simulate emotional desicion-making within an AI?, and (b) How do you program an AI to detect emotions? – Tariq Ali Apr 10 '16 at 15:02

I submitted the questions, but don't expect them to get upvoted a lot. I guess the hope would be that if this question is raised again, the moderators of the eventual site would not close this question but edit it to narrow its scope to more acceptable levels. – Tariq Ali Apr 12 '16 at 0:11

show 1 more improvement suggestion

up vote 13 down vote

added by Blubberguy22 Jan 18 '16 at 20:15

Perhaps "What are the detailed requirements of the "Turing Test?" – kevlarjacket Mar 14 '16 at 16:07

up vote 11 down vote

added by Hazem A. A. Nomer Jan 25 '16 at 14:49

up vote 11 down vote

added by Marqin Feb 15 '16 at 11:00

It depends on what you understand under consciousness. What "Chinese room" actuall do, it forces you to think about your understanding ofconsiousness. As for me I don't see any problems with considering "Chinese room" as conscious subject. – Sklavit Apr 23 '16 at 13:02

up vote 11 down vote

added by x-x Mar 19 '16 at 17:35

up vote 10 down vote

added by Eric Platon Dec 15 '15 at 9:35

up vote 10 down vote

added by Eric Platon Dec 15 '15 at 9:37

up vote 10 down vote

added by Eric Platon Dec 15 '15 at 9:40

up vote 10 down vote

added by Iter Ator Dec 26 '15 at 12:46

In the light of this discussion, do you think this kind of question better belongs here or in the DataScience pool? Answering this should define a clear boundary between both. – Eric Platon Dec 28 '15 at 1:24

Without this type of questions I don't think that an AI site would have enough topic to work as a QA site – Iter Ator Dec 28 '15 at 14:05

up vote 10 down vote

added by kenorb Dec 29 '15 at 11:20

up vote 10 down vote

added by kenorb Dec 29 '15 at 11:27

up vote 10 down vote

added by kenorb Dec 29 '15 at 11:29

up vote 10 down vote

added by kenorb Dec 29 '15 at 12:11

to broad? this question is incredibly variable. – Raven Jan 2 '16 at 0:28

Perhaps not too broad. History shows several attempts and some convergence (but I do not know recent achievements). However, the answer risks to evolve over time, until convergence is really stable :-P – Eric Platon Feb 20 '16 at 1:08

True. the answer might be more of "some approaches include: a... b... c... " instead of a detailed technical explanation. – Raven Mar 24 '16 at 5:10

up vote 10 down vote

added by kenorb Jan 3 '16 at 12:46

Is this question best here or in DataScience? It may just be a question of deciding where once and for all, but answering this question sounds like a prerequisite for an AI community to take off. – Eric Platon Jan 4 '16 at 23:45

@Eric: Probably it overlaps. It can be generalized to "How can AI be trained to recognize stored training data becoming obsolete." – SF. Mar 30 '16 at 12:25

up vote 10 down vote

added by Franck Dernoncourt Jan 6 '16 at 3:03

up vote 10 down vote

added by Franck Dernoncourt Jan 6 '16 at 3:04

up vote 10 down vote

added by Franck Dernoncourt Jan 6 '16 at 3:04

up vote 10 down vote

added by Matt Clark Jan 7 '16 at 4:02

up vote 10 down vote

added by Chase Cromwell Jan 12 '16 at 22:39

up vote 10 down vote

added by DanTheMan Jan 13 '16 at 4:03

up vote 10 down vote

added by maazza Jan 19 '16 at 15:47

up vote 10 down vote

added by Eric Platon Feb 4 '16 at 1:09

Related phrasing: "Is it possible to build a propositional or predicate logic engine with neural networks?" – Eric Platon Feb 12 '16 at 8:47

up vote 10 down vote

added by Eric Platon Feb 12 '16 at 8:48

up vote 10 down vote

added by John Conor Cosnett Feb 13 '16 at 21:47

is that "artifcial" intelligence or biology ? – maazza Feb 16 '16 at 15:36

@maazza If we are trying to build an information processing system "like" the human brain we might want to know a bit about the system that already exists. – John Conor Cosnett Feb 17 '16 at 18:03

That is a good question. Biology is off topic here. But in my opinion we have to include the brain's computational principles in the conversation. After all the connectionist machine learning paradigm (Geoffrey Hinton et al) is heavily inspired by the brain. – John Conor Cosnett Feb 17 '16 at 19:54

How do you think about reformulating the question? I for one agree with you, but I'd vote up a question that shows explicitly the relationship to AI. Perhaps something like "How do AI technologies relate to how the human brain stores data?". Well, this is wordy, but that's about the idea. – Eric Platon Feb 20 '16 at 1:05

up vote 10 down vote

added by Marqin Feb 15 '16 at 10:57

up vote 10 down vote

added by user Mar 2 '16 at 13:03

up vote 10 down vote

added by ferrybig Mar 3 '16 at 9:42

up vote 10 down vote

added by x-x Mar 13 '16 at 23:06

why the down vote? – x-x Mar 17 '16 at 22:47

Indeed, I also would like to know why this question has been down voted (only once). As a matter of fact, several books on AI show quite some connection to nano-technologies and biotechnologies. – Eric Platon Mar 31 '16 at 0:03

up vote 10 down vote

added by Programmix Mar 20 '16 at 20:05

up vote 10 down vote

added by pkhlop Mar 22 '16 at 16:11

up vote 10 down vote

added by ByteCommander Mar 23 '16 at 9:09

up vote 10 down vote

added by Raven Mar 24 '16 at 5:14

up vote 10 down vote

added by pkhlop Mar 25 '16 at 0:36

up vote 10 down vote

added by Omniwombat Mar 25 '16 at 0:42

up vote 10 down vote

added by Omniwombat Mar 25 '16 at 0:48

up vote 10 down vote

added by ratchet freak Mar 29 '16 at 10:51

up vote 10 down vote

added by Tariq Ali Apr 10 '16 at 16:30

	Frustration: Fast duplicated input? :) – nsandersen Apr 16 '16 at 14:57
	Boredom: Repeatedly inputting nonsensical input, such as "What would wood do?" – wizzwizz4 Apr 27 '16 at 19:14
	When I proposed the question, I was imagining the answer to only refer to visual identification of emotion (such as Microsoft's "Emotion API"). Thinking about how chatterbots could identify emotional states would be a very interesting problem and one that I haven't thought about. I'm actually wondering whether this question is too vague, since there can be many different types of AIs, each being able to detect emotions in multiple ways. – Tariq Ali Apr 27 '16 at 22:47