Area 51 » Science

264

followers

This proposal is now in the Beta phase — example questions are locked!

81 Example Questions (4 closed)

active newest votes

up vote 6 down vote

added by Tariq Ali Apr 10 '16 at 16:12

up vote 4 down vote

added by SF. Mar 30 '16 at 12:20

example: locating square and rectangular licence plates on photos of cars. – SF. Mar 30 '16 at 12:21

up vote 3 down vote

added by JeD Mar 20 '16 at 17:50

up vote 3 down vote

added by KKZiomek Apr 28 '16 at 2:54

up vote 2 down vote

added by nsandersen Apr 16 '16 at 14:53

(I think it is a question of how complicated code/logic and at what future point.) – nsandersen Apr 16 '16 at 14:54

up vote 1 down vote

added by James Jan 7 '16 at 18:37

closed as not a real question by Robert Cartaino Apr 8 '16 at 15:29

It's difficult to tell what is being asked here. This question is ambiguous, vague, incomplete, overly broad, or rhetorical and cannot be reasonably answered in its current form.

up vote 1 down vote

added by dorien, edited by Gwenn Apr 24 '16 at 6:48

closed as duplicate of What are the advantages of using a convolutionary neural net versus boltsman machines? by Surb, Robert Cartaino Apr 24 '16 at 14:32

This question covers exactly the same ground as another question in this proposal.

up vote 1 down vote

added by x-x Mar 31 '16 at 1:52

closed as not a real question by Robert Cartaino Apr 8 '16 at 15:29

It's difficult to tell what is being asked here. This question is ambiguous, vague, incomplete, overly broad, or rhetorical and cannot be reasonably answered in its current form.

up vote 1 down vote

added by rcpinto Apr 4 '16 at 0:41

up vote 1 down vote

added by Nathan Apr 6 '16 at 23:22

up vote 1 down vote

added by rnrneverdies Apr 7 '16 at 23:39

up vote 1 down vote

added by Matt Carmona G Apr 15 '16 at 2:20

up vote 0 down vote

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

up vote 0 down vote

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

up vote 0 down vote

added by user Mar 2 '16 at 12:58

closed as duplicate of Is it possible to train AI to solve math problems? by Robert Cartaino Apr 8 '16 at 15:32

This question covers exactly the same ground as another question in this proposal.

up vote 0 down vote

added by Raven Mar 24 '16 at 5:20

up vote 0 down vote

added by Tariq Ali Mar 28 '16 at 0:21

up vote 0 down vote

added by dorien, edited by wythagoras May 5 '16 at 8:42

up vote 0 down vote

added by Alberto Apr 2 '16 at 16:23

up vote 0 down vote

added by rcpinto Apr 4 '16 at 0:39

up vote 0 down vote

added by rcpinto Apr 4 '16 at 0:40

up vote 0 down vote

added by rcpinto Apr 4 '16 at 0:42

up vote 0 down vote

added by SF. Apr 5 '16 at 9:43

up vote 0 down vote

added by enedil Apr 9 '16 at 19:30

up vote 0 down vote

added by lucvoo Apr 10 '16 at 17:45

up vote 0 down vote

added by ratchet freak Apr 12 '16 at 13:24

I'll grant that "complex" is under-defined here. Especially because how little you need for turing completeness. – ratchet freak Apr 12 '16 at 13:25

Maybe fix the question to be "How large would a neural net need to be to be turing-complete?" – Tariq Ali Apr 13 '16 at 13:02

up vote 0 down vote

added by wizzwizz2 Apr 26 '16 at 17:43

You could use a machine-learning algorithm to interface with the NAO robot, I suppose, but the NAO robot is just a robot. Any software running on it is not the NAO robot, but is a program interfacing with the NAO robot. aldebaran.com/en/cool-robots/nao – wizzwizz4 Apr 26 '16 at 18:03

On the other hand, many people use the term "robot" to refer to software and algorithms without any physical form (see "chatterbots" as an example). Those software and algorithms can be said to be 'artificial intelligence'. It's possible that people "assume" that if an algorithm is intelligent, it is therefore a bot. But probably not the other way around (an actual "robot" is not intelligent). – Tariq Ali Apr 28 '16 at 12:15