Talk:intelligent system

RFV discussion

This entry has survived Wiktionary's verification process.

Please do not re-nominate for verification without comprehensive reasons for doing so.

---

Really? Or is it spam for "Intelligent Systems Corporation". SemperBlotto 22:19, 12 January 2010 (UTC)

"Intelligent Systems" is a term widely used by the Computational Intelligence community and should not be mixed with "Intelligent Systems Corporation", who should not highjack the scholarly entry "Intelligent Systems". — This comment was unsigned.

Please provide citations for use in the sense given. Usually, the wordier the definition the harder it is to find citations and show that the citations actually support all the elements of the definition. The definition given is not one I would want to have to provide citations for. DCDuring TALK 22:33, 12 January 2010 (UTC)

Had to change the def, sorry, but it was rubbish. Also I'm doubtful of the definition of "computational intelligence" - Wikipedia says it is an "offshoot" of AI, differentiating it from GOFAI, but not from AI. If it means to include the study of types of intelligence (or problem solving) not inspired by emulating human thought, it is still not separate from AI. Pingku 18:59, 28 January 2010 (UTC)

What are "SC techniques" and how does that establish that AI techniques are involved? What basis is there for the "particularly" clause. DCDuring TALK 20:21, 28 January 2010 (UTC)

(1) I'm going with "Simulation and Control", from the book title (whence also NDS). (It appears to be real - see, e.g., [1].) It's not obvious from the grammar, but what they're describing implying is searching the "solution space" of potential models to find the optimal model. Searching solution spaces, and the various techniques for doing so, pretty much sums up what classical AI is about. (Reading it this way, I don't think it really matters what "SC" means.)

(2) What I want to encompass is that a system: (a) may have a design/purpose that requires AI techniques to work well or perhaps at all; or (b) may use AI techniques throughout but not in a way that is critical to its purpose. For example, the design of an advanced German tutor might require a model of how each student responds to different teaching methods and how advanced he/she is, and a decision process about how then to proceed. A game program might use AI techniques in many places to improve efficiency or to make the game more interesting, but still be much the same game without them. I suggest that the German tutor has the better claim to being called an "intelligent system", but both might be so called.

Cheers, Pingku 12:47, 29 January 2010 (UTC)

The overall approach you are taking seems prescriptive/normative. I know it is more work to inductively extract meaning-in-use from a variety of usage instances, but users expect us to have done something like that. Including a reference to AI may serve users interested in the history of such matters, but introduces encyclopedic elements into the definition and makes it harder to cite, it seems to me.

Also a citation that has elements that even the contributor doesn't understand ("SC") seems a poor choice to appear even on the citations page, let alone in principal namespace. Could SC refer to, for example, Statistical Classification, which is not really an AI technique? I wonder whether a definition as narrow as you are suggesting and attempting to cite won't require a narrow context, like "artificial intelligence". DCDuring TALK 16:38, 29 January 2010 (UTC)

Hmmm. It seems SC refers to Soft Computing (fuzzy logic, neural networks, genetic algorithms and such), which from the early 90's has been part of AI research. SC provides "new" ways to search the "solution space", ways that can be faster but also may produce non-optimal solutions. Ugh! More encyclopaedic stuff. Pingku 19:11, 29 January 2010 (UTC)

The sociology of academe? The old-line AI folks (MIT, Stanford, CMU, et al} vs newer schools of thought? "Soft Computing" might have enough adherents publishing books and articles to warrant an entry. Does SC have its own associations, conferences etc? Or is it just someone's umbrella for various techniques somewhat distinct from the older AI paradigms? DCDuring TALK 23:28, 29 January 2010 (UTC)

There are numerous books on the topic, including the book series "Advances in Intelligent and Soft Computing" (originally "Advances in Soft Computing") and "Studies in Fuzziness and Soft Computing". The International Conference on Artificial Intelligence and Soft Computing will be held in Poland this year, and has been going at least since 2002. There is also a journal (Journal of Multiple-Valued Logic and Soft Computing), which appears to be long-standing (volume 9 was published through 2003, according to the website). Pingku 17:04, 30 January 2010 (UTC)

Seems to pass? — Beobach 22:19, 20 November 2010 (UTC)

• Verified – move to RFD if you're still not convinced. Ƿidsiþ 08:44, 28 March 2011 (UTC)