Where knowledge is beyond a machines "knowledge": taxonomy classifications

A machine had to believe a classification (de) if it were able to do something like believing. That is because of the absence of knowledge which would enable the machine to judge what notion is a narrower than another: There is no such knowledge. There are only a assignments between broader and narrower terms and associative relationships etc. But the knowledge a machine would need to verify those assignments is missing: A machine were able to verify the assignment, if it would know how far//by what broader and narrower term are similar and where they differ. – A classification could provide this data. That would be an other assignement. But if a machine had its own data structure providing this kind of data – would that be knowledge or just assignments as well?<<

---

[locally referred by: ./. ]

Just found Walter Smits' note about his work on this topic

Just found Walter Smits' note about his 2002 work on this topic. Possibly it would be useful to get in contact with him. From his note:

About my study: someone *had* to ask... :-s ;-). Very roughly spoken, when I've finished this study I should more or less be able to put the knowledge that someone's got in his brains into a computer. One strange property of the human being is that we often know thinks we don't know about. We've got the knowledge, we use it, be we don't know about it (it's become some kind of automated behaviour).

---

[locally referred by: ./. ]

Some books are in reach to me

[Bearing:] Dreyfus' book What computers can't do and Baum's What is thought? (referred to in Wikipedia's computer program entry) are in reach to me in my local university library.<<

---

[locally referred by: ./. ]

Dreyfus, 1972: "it is not possible to capture expert knowledge in an algorithm"

Just in my very first attempts to draft the main idea of my theses I wanted to know why people refer to a computer program (de) as a knowledge storage. Since Wikipedia (de) didn't give much of help there I quickly searched the web. I found:

"Dreyfus illustrates his claims with references to the problems faced by AI researchers who attempted to codify expert knowledge into computer programs. The success or failure here really has little to do with the computing machinery, but with whether expert competence in the domain in question can be captured in an algorithmic procedure. In certain well-circumscribed domains this has succeeded; but more often than not, argues Dreyfus, it is not possible to capture expert knowledge in an algorithm, particularly where it draws upon general background knowledge outside the problem domain."   (source)

<<

---

[locally referred by: ./. ]

And now to something completely different

I changed my mind. It would be of much more use to publish this work in English. More people can participate and discuss the theses I am working on. I want to set up a development website on a web based free software developing platform (like alioth or Source Forge) as soon as possible since I am seeing neurology is meeting my place coming from the natural science side. Since I am working on this since about 2001 I am not willing to look like a copycat when the neurologists are done with the main job of "disassembling" brain, i.e. finding out how it is basically working like. Hence I want to publish my theoretical work prior to theirs based on natural science. (Mine is based upon philosophy/human sciences, software development experiences, and much interest about how thinking works.)

This blog is about to switch to a research and preparation blog, collecting citations from the web usable to underpin my theses, and for drafting pages I am going to set up on the software development site.<<

---

[lokal referenziert von: ./. ]

Alte Beschreibung dieses Blogs

Dieses Blog wurde vor dem Wechsel ins Englische wie folgt beschrieben. Vielleicht kann der erste Eintrag dieses Blogs einen Eindruck davon vermitteln, wozu dieses Blog ursprünglich bestimmt war.

Innerhalb von FAVA ist ein Begriff anhand seiner Eigenschaften wiederauffindbar. Die Precision einer Suche könnte dadurch erheblich steigen. Gerade für Laien bietet sich daher an, FAVA anstelle eines herkömmlichen Begriffssystems einzusetzen, etwa für Webseiten, MP3s... Erweitert zu einer Peer-To-Peer-Plattform sollte sich ein kooperatives, verteiltes Retrieval-Instrument bilden lassen.<<