Talk:Computer Science:Artificial Intelligence

From Wikibooks, the open-content textbooks collection

Guidelines for contributors

Welcome!

Please don't add a bunch of empty sections unless you will fill them in later. Nobody likes to be given writing assignments.

I like the big buckets first approach, so I'm not going to mess with subpages or namespaces until this page is really fat.

--Waxmop 00:49, 27 Jan 2005 (UTC)

This book needs to be improve. First, it talks about state space search. Then a section about EMAC and LISP is presented. The links between the sections and the sections are not well explained. Pfv2 02:51, 3 October 2005 (UTC)

When I took AI at my high school, we used Python rather than LISP, and I believe that Python is beginning to supplant LISP as the AI language. Therefore, shouldn't we add at least discuss Python as well? IMacWin95 15:19, 18 March 2006 (UTC)

[edit] Yes, it needs some love.

I typed in the first half of my class notes, then my son was born, and I haven't had time to finish. Please feel free to make whatever changes you like. --Waxmop 03:44, 19 October 2005 (UTC)

I say it's about time to start chopping everything up into well thought out chapters. Considering the size of the subject, this is going be a very large book (if it takes off). It would probably be best to think the structure through thouroughly before going any further with it.

I think it should start with the basic necessities like programming languages (Prolog & Lisp, these could just be simple introductions. There are wikibooks if the reader wants to learn the language itself), Logic (Propositional and Predicate Logic) and basic stuff like state space searching and simple gameplay. It could then move on to more advanced topics like Natural Language, Knowledge Representation and Machine Learning. It would probably be good to brainstorm for a while to see what needs to go where. For instance, probability theory could get it's own section, or it could be a part of the Machine Learning section. I personally favor as much modularity as possible, so I'd go with the first opstion, but that may not be possible for all subjects. It might also be nice to go into the psychological/philosophical side of things, to make it more than the standard technical overview of the field.

risk 14:22, 20 October 2005 (UTC)

[edit] conspicuously missing; Psychology cross discipline

I'd like to say that the articles i just skimmed looked like they were pretty good standard material for explaining hardware and software issues.

Whats obviously missing in my mind are the models of human intelligence which are then remodeled in code.

I'd be interested in how and where you might feel that such an exploration would fit inside of your current structure, or, what you might envision for the interdisciplinary crossover structure.

I might add that as an Aspergers Syndrome dyslexic, math as a child was always presented to me via serious cognicentrism, and i never found good ways to compensate like i did for verbal information, with the possible exception of my handling of fractions.

Thus, while i know that i am patently useless in the mathematics arena, I have contemplated AI from the limited standpoint of coding brain region functions in "BASIC". (For instance.)

209.129.49.65 00:26, 3 February 2006 (UTC)

Prometheuspan 00:26, 3 February 2006 (UTC)

Whats obviously missing in my mind are the models of human intelligence which are then remodeled in code.

It's because we don't have such a model yet. (Or more precisely, we do have a lot of totally different models, which model different parts of intelligence, rather poorly. ) The connectionist (neural network) model for example needs so much Math that it's even hard to understand for people who don't have Asperger.) The most comprehensive book about AI is the AI: A modern approach by Peter Norvig and Stuart Russel. It has some philosophical aspect about AI, doesn't require too much math (except the neuaral network and probability network parts), but requires so much logic, that it's practically math. If you want to learn more about human intelligence without dvelving too much in math, you should probabily read some book about cognitive psychology.

Anyway, as I see this wikibook is rather for complementing the Norvig book, which I very much welcome, because the Norvig book contains only (often rather hard to comprhend) pseudocode. --Zslevi (talk) 11:04, 12 May 2008 (UTC)