Portal talk:Artificial intelligence

From Wikipedia, the free encyclopedia
Jump to navigation Jump to search
WikiProject Portals  
WikiProject iconThis page is a portal. Portals are within the scope of WikiProject Portals, a collaborative effort to improve portals 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 portal has not yet received a rating on the project's quality scale.
 ???  This portal has not yet received a rating on the project's importance scale.
 
Note icon
See also: Instructions • Guidelines • List of Portals
Note icon

Please take care when editing, especially if using automated editing software. Learn how to update the maintenance information here. (July 2018)
WikiProject Technology (Rated Portal-class)
WikiProject iconThis page is within the scope of WikiProject Technology, a collaborative effort to improve the coverage of technology 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.
 Portal  This page does not require a rating on the project's quality scale.
 
WikiProject Computing / Software / CompSci (Rated Portal-class)
WikiProject iconThis page is within the scope of WikiProject Computing, a collaborative effort to improve the coverage of computers, computing, and information technology 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.
 Portal  This page does not require a rating on the project's quality scale.
Taskforce icon
This page is supported by WikiProject Software.
Taskforce icon
This page is supported by WikiProject Computer science.
 

Portal motivation[edit]

Due to the expansion of the original article on AI and the heavy debates on the topic this requested portal was created. A redistribution of the text on the original article is still called for. The aim with this portal is to provide an objective index to the various independent aspects of AI. Additionally I tried to replace or repeat as little as possible of the original text, but still create a meaningful entry page to the topic. --moxon 22:12, 10 October 2005 (UTC)

Article:Aritificial intelligence[edit]

For a discussion history of the topic refer to Talk:Artificial intelligence. Because of the extended amount of references to the original page it is suggested that the referensing articles not be changed. It is further suggested that the original title not be redirected to the portal, but rather a very short explanation of the subject be given, as would be reported by Google, Answer.com, etc. --moxon 03:08, 11 October 2005 (UTC)

ai-stub to get cancelled[edit]

Has anyone noticed there's a move to cancel the ai-stub class of stub. Surely there's a bunch of stuff that could be stubbed. See Wikipedia:stub types for deletion. --Jaibe 19:22, 5 July 2006 (UTC)

As of today, the "Compu-AI-stub" is in use. See the page Category:Artificial intelligence stubs for a list of AI stubs. That page could perhaps serve as a to-do list until we such a list onto the AI portal itself. EverGreg (talk) 14:18, 16 January 2008 (UTC)

Comments on the page content: Machine learning[edit]

Machine learning is IMHO not associated with conventional AI, but with the more modern, statistical approach seemingly termed computational intelligence on the page. To convince yourself of this, think of the journal JMLR and its content, or the conferences ECML and ICML. --Scellus 13:44, 28 September 2006 (UTC)

New here[edit]

What should I do? - Peregrine Fisher 09:47, 5 January 2007 (UTC)

Computer Science/Film student taking course in AI and looking for project wants to help write an article about AI. What can I do? Cobaltghost (talk) 17:54, 26 November 2007 (UTC)

Your basic choices are:
  1. Try to find a notable subject that hasn't been written about. (But do a thorough search on the subject in wikipedia first: with 2M+ articles, many subjects have already been covered. Poke around the category hierarchy to get a feel for what's there.)
  2. Read through articles on subjects you know well and try to identify the BS and replace it. One clue: if the article doesn't have footnotes[1] or harvard references (LikeThis 1908), there's probably some paragraphs that are off-topic or just wrong.
In either case, you should research your subject carefully. Make sure you get the books and articles you will need. Its also worth it to read Wikipedia:Guide to writing better articles, before you edit an existing article or create a new one.
As for a specific suggestion, with your film background, you might be interested in becoming an expert on artificial intelligence in film. Look into Category:Artificial intelligence in fiction and the article artificial intelligence in fiction. Another area that combines media and computer science would be Category:Multimodal interaction (which includes things like virtual reality) ---- CharlesGillingham (talk) 21:54, 26 November 2007 (UTC)
Still another more general suggestion, especially for anyone taking a course in AI or an AI related field. As you read your text, check the articles on the same topics in Wikipedia, and help make them more accurate and mark them with references to the page numbers in your text. Use Wikipedia as your class notes. Most of the articles in Wikipedia have sentences, paragraphs and entire sections that are just plain wrong. We are in desperate need of more accurate sources. ---- CharlesGillingham 08:09, 1 December 2007 (UTC)

Notes[edit]

  1. ^ Like this

Article request: Inceptionism[edit]

Hi, someone please create the article for Google's deep image invertions. They call it "Inceptionism" so that could be the name of the article - alternatively (and possibly preferrably) one could use a more explanatory title.

It is explained here: http://googleresearch.blogspot.co.uk/2015/06/inceptionism-going-deeper-into-neural.html (also includes 4 research papers) and has been picked up by countless news-agencies.

If this is the wrong place to ask about this please forward it to people who might create this article or redirect me there. Thanks.

--Fixuture (talk) 16:11, 20 June 2015 (UTC)

Okay the article exists now: DeepDream --Fixuture (talk) 17:11, 27 August 2015 (UTC)

Help please - I'm trying to start a list somewhere on Wikipedia of human abilities that will be tough for artificial intelligence[edit]

I am completely new to Wikipedia contribution.

We Wikipedia contributors need to start a list of abilities, qualities, and experiences of humans that will be difficult for an artificial intelligence to have.

Wikipedia already has an entry "Outline of thought"; therefore, the list I propose should be appropriate for Wikipedia proper (as opposed to WikiProjects, WikiBooks, etc.).

I wrote to the help/info email address info-en-o@wikimedia.org.

This is what I wrote:


My question will refer to Portal:Artificial_intelligence

The title of the Wikipedia Portal is Artificial Intelligence.

Please see the section on the right hand side, "Major Problems of AI."

I have an idea that is completely appropriate to be made available at Wikipedia.

The idea is a list of human abilities that artificial intelligence will have trouble with.

This is a similar category to "Major Problems of AI."

The concept I envision is different to it, however, because it will be an extremely long list of specific abilities written in layman's English.

Of course the list should be arranged in categories.

One such category could be "Emotion."

"Major Problems of AI" has some items which inform "Emotion" (Reasoning, Machine Learning, Natural Language Processing, and strong AI), but if one clicks each of those links and reads them, so little of the content pertains to emotion directly.

In addition to "Emotion," here are examples of other items in the list I propose: Creativity, Imagination, Art, and Humor, each of which can be broken down. For example, Art can be broken down into Music, Dance, Literature, Visual Art, and so on. Emotion as an item in the list can be broken down into Anger, Fear, Joy, Sorrow, Humiliation, Pride, and so on.

A section of this list that I propose would be a list of experiences that an artificial intelligence would have trouble with -- for example, what it's like to raise a child, grow old, feel physical pain, feel lazy, feel jealousy (there's overlap with the emotions list here), what it's like to experience a hallucination or illusion, to miss or mourn over a loved one, to experience camaraderie, to lead a group of people, to smell what humans can smell, to experience what petting a dog feels like as a sensation and as a psychological experience, to feel bias, to have or reject religious faith, to deceive a human; other qualia, and so on.

Again, the list will be extremely long. They will be categorized. This corpus of content will be enormous [like the entry "Outline of thought"], which makes it appropriate for Wikipedia. One benefit of it: it will inform designers and theorists of artificial intelligence.

Another benefit is that it will inform the general public and students of the field in an accessible, easily understood way.

Eventually as users peruse the list, they will have to ask questions like, "Must an entity be a wet, mortal, living organism first in order to experience these things?"

Another question arises in the process: "...'intelligent' according to whose definition? Humans'?"

"If yes, perhaps according to average humans? What's an average human? Does an average human speak English? Is an average human from a technologically advanced country?" and so on.

"How similar to a human must the artificial intelligence be in order to satisfy the notion of strong AI?"

"If it is not similar to a human, can it fulfill our notion of strong AI? For example, could it run for public office and engender enough empathy from human voters to defeat a human candidate, when it doesn't know what it's like to be a human in so many ways?"

"For example, if it doesn't know what it's like to smell what humans can smell, or what it's like to feel pain like humans do, much less what it's like to, say, raise a child, it will never understand huge sections of poetry and art; its bonding and conversing with humans in person will be sub-par, and it will alienate us, surely if we expect a leader to nurture subordinates and voting constituents."

This wikipedia list I propose will also go into what progress in the field of artificial intelligence has been made in each item in the list -- and future directions.

I do not know where in the large wikipedia universe the list should be started, but I would like it to be in wikipedia proper, as opposed to wikibooks or wikiProjects, for example. I tried to figure out where, but I am completely new to contributing to wikipedia, and wikipedia is still not user-friendly enough for beginning contributors, and is divided into so many parts that are similar in content and purpose to one another.

I am almost certain that this message has to be forwarded to a specialist at wikipedia knowledgeable in artificial intelligence enough to grasp the virtues of the wikipedia list I am proposing, please.


I received a reply:

This is a content issue that is beyond the scope of this service, however that portal has an associated "Talk" page where you can add this suggestion.


Hence I wrote to them again asking them to forward it to the appropriate content moderator.

Yes, my email got carried away, but I do think the list will be useful to anyone who believes that strong artificial intelligence is possible, especially those who believe it will occur within decades.

The entry could be titled the way I wrote it up top, or "Future limitations/problems of AI." That section "Major Problems of AI" exists, but the links in it don't suffice.

Also, the list will be of use to science fiction writers and anyone who writes about artificial intelligence, regardless of whether they're researchers.

I will try to start the list, but I have never created an entry in wikipedia -- this talk is the first time I've "contributed."

If anyone else wants to and knows how, etc., that'd be great.

-Nicholas Nn9888 (talk) 08:30, 16 July 2015 (UTC) Nn9888 (talk) 07:38, 14 July 2015 (UTC)

Hello Nicholas. There is not a complete consensus within the field about what is impossible, and this is large and complicated topic. I would suggest reading all these articles, and you will begin to understand the issue from the inside: Philosophy of artificial intelligence, History of AI (especially History of AI#The problems), Moravec's paradox, Chinese room, Turing test, "Computing Machinery and Intelligence", Dreyfus' critique of AI, and (of course) AI, especially AI#Approaches. Good luck! ---- CharlesGillingham (talk) 20:39, 7 September 2015 (UTC)

Category:Open source artificial intelligence[edit]

I just created Category:Open source artificial intelligence, could you help out with it please?

As there's still no WikiProject for artificial intelligence I'll just post this here.

--Fixuture (talk) 15:02, 4 January 2017 (UTC)

Status report from the Portals WikiProject[edit]

Wikipedia:WikiProject Portals is back!

The project was rebooted and completely overhauled on April 17th, 2018.

Its goals are to revitalize the entire portal system, make building and maintaining portals easier, and design the portals of the future.

As of April 29th, membership is at 56 editors, and growing.

There are design initiatives for revitalizing the portals system as a whole, and for each component of portals.

Tools are provided for building and maintaining portals, including automated portals that update themselves in various ways.

And, if you are bored and would like something to occupy your mind, we have a wonderful task list.

From your friendly neighborhood Portals WikiProject.    — The Transhumanist   03:22, 30 April 2018 (UTC)