Stack Exchange
log in sign up chat discuss faq users badges

Area 51 » Technology

definition commitment beta

The AI site didn't have enough activity during the beta, and has been closed. You can download the data dump of all questions here.

274
followers

Artificial Intelligence

Q&A site for theorists, system architects and analysts of intelligent machines and software.

9
votes
4
answers

Faked Artificial Intelligence like in Game Development

aug 10 '16 at 11:08 kenorb 4,638
11
votes
4
answers

How is this proposal different from Cross Validated?

jun 30 '16 at 8:51 Community♦ 1
7
votes
2
answers

Artificial Intelligence proposal scope

jun 30 '16 at 8:51 Community♦ 1
18
votes
2
answers

What niche does this AI proposal aim to fill?

jun 30 '16 at 7:59 Community♦ 1
24
votes
7
answers

What will prevent this proposal from going down again?

jun 28 '16 at 18:16 Community♦ 1

discuss this proposal

70 Example Questions (2 closed)

active newest votes
up vote 16 down vote
What kind of learning algorithm is suitable for recognizing motion with a camera?
added by echad Jul 22 '13 at 17:04
link
up vote 13 down vote
How can I design a neural network to handle a variable number of inputs?
added by trav1s Jul 17 '13 at 8:39
link
up vote 12 down vote
Is there any research into machines automatically "learning new ways of learning"?
added by SF. Jul 22 '13 at 11:08
link
1  
not just gathering/storing one class of data indefinitely, but actually expanding scope of search basing on data received, e.g. able to turn a set of rules of a syntax into a syntactic parser, and then use said parser to parse another set of rules - written in syntax of the previous parser, to generate another, still usable one? – SF. Jul 22 '13 at 11:21
I was also thinking about this recently, but the way I framed it was, "what if the learning parameters were adapted and learned? Would this be more efficient for some classes of problems?" – trav1s Jul 23 '13 at 3:03
1  
This is phrased similarly to a reference request as allowed on the math SE. I think reference requests are OK, but not the greatest questions. It is kind of like asking for someone else to search the literature for you. – trav1s Jul 23 '13 at 3:07
@trav1s: It is a minute of answer for an expert, and a week of search for a layman. Especially if the answer is "no, there is none" the search may take a long, long time. – SF. Jul 23 '13 at 6:38
I agree with that, but the questions are better to be from the expert than from the layman. – trav1s Jul 23 '13 at 7:47
up vote 12 down vote
Why doesn't my neural network learn?
added by BartoszKP Sep 18 '13 at 13:12
link
3  
A very common question on SO site, with many different answers and scenarios. Probably such questions should belong here. – BartoszKP Oct 16 '13 at 12:43
up vote 11 down vote
What is the meaning of Fuzzy Logic and where is it applicable?
added by Othman A.A. Jul 24 '13 at 13:08
link
up vote 11 down vote
What can a super-Turing machine do that a Turing machine can't?
added by trav1s Aug 1 '13 at 3:02
link
up vote 11 down vote
Which algorithms exist for learning patterns from images and identifying these patterns in other images?
added by Quonux Aug 1 '13 at 12:52
link
up vote 11 down vote
Is choosing the best kernel method for your data set just trial and error?
added by nijansen Aug 6 '13 at 11:16
link
up vote 11 down vote
Does a machine learning algorithm exist, which is able to extract profitable patterns from a forex timeseries?
added by Maecky Aug 25 '13 at 16:10
link
up vote 11 down vote
What are the differences in implementation of the A* algorithm when the state space is either a tree or a graph?
added by Luiz Vieira Oct 26 '13 at 23:04
link
up vote 10 down vote
When should hard-coded signal processing be used vs adaptive processing or a combination of both?
added by trav1s Jul 17 '13 at 3:07
link
up vote 10 down vote
Is there a grand, unified theory of artificial intelligence?
added by e-sushi Jul 26 '13 at 11:14
link
up vote 10 down vote
Is a fuzzy control system the only AI approach for solving control system problems, like the inverted pendulum?
added by echad Jul 26 '13 at 11:36
link
up vote 10 down vote
Are neural algorithms or more computational algorithms better suited for face recognition?
added by Quonux Jul 28 '13 at 0:13
link
6  
"better" is ambiguous. This question lacks precision, however the topic is interesting. Maybe it would be helpful to choose a particular criterion. – BartoszKP Aug 17 '13 at 21:45
up vote 10 down vote
Does it make sense to combine fuzzy logic and the Bayes theorem when implementing self-learning pattern recognition functionalities?
added by e-sushi Jul 28 '13 at 11:10
link
up vote 10 down vote
Why does adding increasingly more neurons lead to overfitting errors? Wouldn't more processing power lead to better results?
added by trav1s Aug 1 '13 at 2:43
link
up vote 10 down vote
When and how are memristors effective at emulating neurons?
added by trav1s Aug 1 '13 at 2:57
link
up vote 10 down vote
How close are we to passing the Turing test?
added by Keshav Srinivasan Aug 8 '13 at 15:16
link
3  
Good question for philosophers and science fiction fans, but in general completely non-answerable. What answer do you expect? "2 years", "really really close", "we're very far", "10 kilometers"? In fact this asks to predict the future, and also to write almost a book chapter defining the precise version of Turing test one has in mind and what exactly does "passing" mean. – BartoszKP Nov 6 '13 at 23:03
@BartoszKP: Quite answerable if we're to create a metric based on a standarized test and extrapolate. Test many machines against a statistically significant number of people. Each machine will have a different success ("recognized as human") rate. Distance between current best rate and 50% (machine passes for human better than real human) is one simple, objective "how close we are". Create a graph of machine creation date vs success rate, find trend curve. Extrapolate the curve, find point where success rate exceeds 50% and you have a less reliable but easier to grasp time estimate. – SF. Dec 2 '13 at 9:16
up vote 10 down vote
Is a multiagent system program an AI-based multithreaded program?
added by echad Aug 15 '13 at 14:49
link
up vote 10 down vote
Is there a common unsupervised Artificial Neural Network for pattern recognition that adds/removes connections, nodes, and adjusts weights internally?
added by Tim Winter Aug 15 '13 at 16:42
link
up vote 10 down vote
Why are n-gram analysis more popular than bag-of-words for spam detection?
added by JDong Aug 24 '13 at 7:26
link
up vote 10 down vote
Under which circumstances would you choose a neural network over support vector machines and vice versa?
added by nijansen Aug 25 '13 at 14:47
link
up vote 10 down vote
Which machine learning algorithms are suitable for time series forecasts?
added by Maecky Aug 25 '13 at 15:46
link
up vote 10 down vote
How can I achieve reliable model training, i.e. estimates of model parameters, given only little amount of training data?
added by Nikolaus Aug 25 '13 at 17:56
link
up vote 10 down vote
How can I state the significance of an evaluation result, e.g. a confidence interval for classification error rate?
added by Nikolaus Aug 25 '13 at 18:05
link
up vote 10 down vote
What is the exact difference between DBM (Deep Boltzmann Machines) and DBN (Deep Belief Network, also based on RBMs)?
added by lejlot Aug 28 '13 at 15:26
link
up vote 10 down vote
Does it make sense to learn parameters of another learning algorithm?
added by danijar Sep 3 '13 at 20:44
link
This is pretty similar to this other question area51.stackexchange.com/proposals/57719/… – trav1s Oct 29 '13 at 7:12
1  
@trav1s I'd argue that this is a better one, because it's more precise. "learning new ways of learning" seems completely undefined to me. – BartoszKP Nov 2 '13 at 16:03
up vote 10 down vote
Would LISP be a good language to program an AI in? What other languages could this be done in?
added by Jarno Sep 9 '13 at 18:54
link
4  
This kind of question is looking more for a list than a solution to a specific problem. – Adrian Collado Oct 3 '13 at 13:50
What is a "good" language anyway? Is there any language that is not a good language to program an AI in? – Simon André Forsberg Oct 29 '13 at 0:30
up vote 10 down vote
What are all of the possible learning techniques for a neural network? [closed]
added by Jarno Sep 9 '13 at 19:01
link

closed as not a real question by Robert Cartaino Nov 4 '13 at 13:37

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.

5  
All? Do you plan to become a wikipedia/database? There are at least few hundreads of answers for such question. – lejlot Oct 27 '13 at 10:03
1  
This is fundamentally a really broad question and creating this type of all-encompassing list by polling the community for every possible contribution isn't really what we do. See cstheory.stackexchange.com/about – Robert Cartaino Nov 4 '13 at 13:37
up vote 10 down vote
Does supervised learning have any pro's over non-supervised learning, given the huge amounts of information on the internet versus given information.
added by Jarno Sep 9 '13 at 19:04
link
2  
Wording seems rather risky here - this kind of question suggest the answer, which in many cases is simply wrong. Better question would be "What are the pro's and con's ..." then current "Does .... have any pro's" – lejlot Oct 27 '13 at 10:02
up vote 10 down vote
Do all, AI-based, pattern-recognition programs require database(s) ?
added by echad Sep 20 '13 at 14:06
link
1  
[+1] Nice question... will be interesting to read the answers to that. ;) – e-sushi Oct 3 '13 at 22:55
up vote 10 down vote
Would it be more feasible to have a program split a problem into multiple smaller problems for genetic programming or to leave it as one big problem?
added by Adrian Collado Oct 3 '13 at 13:57
link
I assume that such question would be specified to one particular problem in its full form? Such questions do not have one, generic answers, but rather the ones starting with "it depends on many factors" :) – lejlot Oct 27 '13 at 10:06
up vote 10 down vote
How do I choose a classifier for a data set? How will I know when a linear SVM is inappropriate for a problem?
added by JDong Oct 3 '13 at 16:45
link
up vote 10 down vote
What are the possible consequences of changing the K factor in K-fold cross validation? What does it mean in practice?
added by BartoszKP Oct 6 '13 at 9:41
link
up vote 10 down vote
Can linear regression fully replace a perceptron?
added by BartoszKP Oct 8 '13 at 12:54
link
up vote 10 down vote
What algorithms, models, or approaches are used for the AI in <video game/other consumer software>?
added by Cam Jackson Oct 13 '13 at 10:03
link
up vote 10 down vote
Is it a good idea to use artificial intelligence to solve [clearly described problem]? Or is there a classical approach that would solve it better?
added by kdzia Oct 30 '13 at 15:54
link
Good question! Very often someone just uses AI approaches to a problem, just because it's "cool" or something, ignoring the fact that a problem is perfectly tractable by classical, direct methods. – BartoszKP Nov 6 '13 at 10:20
up vote 10 down vote
What are the criteria for calling something artificial intelligence? (as opposed to a tradional algorithm, fake A.I., ...)
added by kdzia Oct 31 '13 at 11:24
link
Cool recurrence :D – BartoszKP Nov 7 '13 at 11:33
up vote 10 down vote
What would be the easiest way for a mathematically challenged person to create an AI program/algorithm to mimic a human personality ?
added by Meterbuchus Nov 9 '13 at 19:36
link
How could a mathematically competent person do it? Mimicking a human personality doesn't sound like a solved problem (unless I misunderstood) – David Robinson Nov 17 '13 at 22:24
Apple's Siri but with a more specific design – Meterbuchus Nov 18 '13 at 20:13
1  
"easiest" - vague ; "mathematically challenged" - vague ; "AI program/algorithm" - vague ; "mimic" - vague ; "human personality" - vague. While it's not a inherently very bad question it seems poorly worded, and has so many possible interpretations that it's extremely broad. – BartoszKP Nov 25 '13 at 11:54
Mathematically Challenged - A lot of people that have a fascination with Artificial Intelligence are not necessarily fluent with the mathematical concepts such as Calculus but want to get involved. – Meterbuchus Nov 27 '13 at 19:17
Yes, I can see them here posting questions, instead of trying the Science Fiction site. Sorry, for being rude. – BartoszKP Nov 29 '13 at 17:35
up vote 10 down vote
What would be the easiest way to integrate Artificial Intelligence into a Botnet ?
added by Meterbuchus Nov 9 '13 at 19:42
link
1  
I'm sorry, what?! – BartoszKP Nov 26 '13 at 16:20
1 2 next

This site has been

Closed

This proposal didn't have enough activity during the beta. It was closed 7 years ago.

You could try proposing a variation of this site that you think will have a better chance.

proposed by

trav1s
10k●5●37●119

7 years ago (edited 7 years ago)

viewed

12,658 times

latest activity

7 years ago

followers

users also following

13.1% Data Science
12.0% Arduino
8.4% Moderators
7.3% Puzzling
20.4% only this proposal

followers active in

69.0% Stack Overflow
19.7% Meta Stack Exchange
14.6% Super User
13.5% Meta Stack Overflow
9.9% Software Engineering
9.5% English Language...

recent followers

added Feb 23 '14 at 9:47
Florian Pommerening
506●1●6
added Feb 19 '14 at 10:50
jkbkot
7,368●26●50
added Feb 18 '14 at 17:05
Wannabesage
51●1
added Feb 18 '14 at 12:55
Xmindz
1,905●1●17●47
added Feb 12 '14 at 0:28
user103280
51●1
added Feb 10 '14 at 23:35
Jens Piegsa
10.3k●8●68●169
added Feb 5 '14 at 11:40
user102883
51●1
added Jan 29 '14 at 12:12
Rhand
993●7●22
added Jan 29 '14 at 10:19
AppleGrew
9,443●24●80●146
added Jan 28 '14 at 18:55
Bharat Kul Ratan
1,365●2●12●30
added Jan 28 '14 at 15:47
icosamuel
812●3●21
added Jan 28 '14 at 13:17
Matt Ellen ♦ 1
51.9k●24●265●507
added Jan 28 '14 at 4:14
disrvptor
1,580●10●25
added Jan 26 '14 at 19:02
Stat-R
12k●20●108●204
added Jan 24 '14 at 13:51
CapelliC
57.6k●5●43●89
added Jan 23 '14 at 15:03
CodesInChaos
157k●32●418●644
added Jan 23 '14 at 11:26
Michael Norris
51●1
added Jan 22 '14 at 18:52
Necro0x0Der
751●1●9
added Jan 22 '14 at 17:16
Martin Ogden
993●4●16
added Jan 21 '14 at 11:57
castarco
2,210●4●23●46
added Jan 20 '14 at 21:03
Victor Stafusa
113k●33●328●598
added Jan 20 '14 at 6:09
MRS1367
1,195●1●13●44
added Jan 16 '14 at 18:30
Malachi ♦ 1
43k●18●180●512
added Jan 16 '14 at 11:47
Mark Smith
51●1
added Jan 15 '14 at 8:34
Natalia
664●1●6●16
added Jan 12 '14 at 22:12
Compukid
1,503●11●43
added Jan 9 '14 at 16:46
Peter Hickman
51●1
added Jan 4 '14 at 6:56
ryanabooth
502●2●9
added Jan 1 '14 at 21:39
Domenico De Felice
1,114●4●26
added Dec 31 '13 at 17:42
Bill B
2,900●19●34
added Dec 22 '13 at 23:28
JavaCake
9,782●15●77●189
added Dec 19 '13 at 4:23
AsianSquirrel
3,280●4●24●66
added Dec 16 '13 at 12:13
xenomancer
161●6
added Dec 15 '13 at 4:01
user2845306
977●10●24
added Dec 12 '13 at 11:33
Yao Bo Lu
851●3●17
added Dec 6 '13 at 17:41
user2883115
61●1
added Dec 6 '13 at 7:06
K.C.
2,632●2●27●53
added Dec 2 '13 at 4:53
ContextSwitch
5,299●8●51●105
added Dec 2 '13 at 2:38
WafflePilot
51●1
added Dec 2 '13 at 1:36
Rob Berkes
51●2
feed icon proposal feed
faq | blog | legal | privacy policy | contact us | feedback always welcome
site design / logo © 2021 Stack Exchange Inc; user contributions licensed under cc by-sa; see the licensing help page for more information.
rev 2021.2.22.577