[Top] [All Lists]

Re: [ontolog-forum] Science, Statistics and Ontology

To: "[ontolog-forum]" <ontolog-forum@xxxxxxxxxxxxxxxx>
Cc: Rich Cooper <rich@xxxxxxxxxxxxxxxxxxxxxx>
From: Ali SH <asaegyn+out@xxxxxxxxx>
Date: Wed, 16 Nov 2011 12:36:03 -0500
Message-id: <CADr70E0gXGBCL8e0wBF7z9BOG5Bd8oBnoQ+3nh-tbm2eFdmuGw@xxxxxxxxxxxxxx>
Dear Rich,

I wrote the bit you quoted.

On Sun, Nov 13, 2011 at 3:21 PM, Rich Cooper <rich@xxxxxxxxxxxxxxxxxxxxxx> wrote:

Dear Len,

I am not entirely sure who wrote the following quote (Len or Simon?) but I would like to replay it for a moment, from the perspective of a learning algorithm which, unattended, senses the environment and is preprogrammed to perform discovery.  Your quote, seen from this perspective, is below:

    If I were to distil the article into three broad themes (incidentally, summarized quite well here http://xkcd.com/882/ ), it'd be:

        1. Misunderstanding the theory behind the statistics

This one seems to be just a math error, unless you mean also that the errors in this category include searching for a logical combination of sensed categories that can be used in discovery.  So unless I misunderstand the point you are making in 1, I consider this category to be fixable in the algorithmic implementation.  Please correct me if you meant something deeper than that. 

This math error has implications in terms of how to combine results and build on a body of knowledge. See the following points.

        2. Incorrectly combining hypotheses (especially from a statistical perspective - but it rests on semantic misinterpretations)

The only combination to be used in my hypothetical discovery system algorithm would be AND, OR and NOT.  All other forms of composition are of course just logical combinations of hypotheses, so it appears that the discovery algorithm would be immune to this category of error as well. 

Not really, though perhaps if your algorithm is working in a silo and not communicating with any other agents then what you say might be adequate. Otherwise, it is necessary for such an algorithm to also specify mappings between the elements of your domain (and the vocabulary used to describe it) and another's. Indeed, this was the point of the critique of the meta-analysis in the linked to article. The logical connectives are almost a non-sequitor.

        3. Incorrectly revising hypotheses

Again, an algorithm that has been preprogrammed to revise hypotheses in the face of sensed evidence can only try one at a time of the various possible revisions, all constrained to be logically correct.  So this error also I find would be eliminated in a discovery algorithm that was properly preprogrammed. 

What am I missing in your opinion?

My description was aimed at scientific discovery within a collaborative context. One where it is not just one agent and one POV, but a group of agents trying to coordinate their vocabularies and build a shared body of knowledge. So the discovery algorithm would have to be able to dynamically generate semantic mappings between the vocabularies deployed by each of the agents whose work you wish to reuse and/or contribute to.

It's not just about logically combining statements, but also about semantically integrating different conceptualizations in order to be able to meaningfully combine results.


Message Archives: http://ontolog.cim3.net/forum/ontolog-forum/  
Config Subscr: http://ontolog.cim3.net/mailman/listinfo/ontolog-forum/  
Unsubscribe: mailto:ontolog-forum-leave@xxxxxxxxxxxxxxxx
Shared Files: http://ontolog.cim3.net/file/
Community Wiki: http://ontolog.cim3.net/wiki/ 
To join: http://ontolog.cim3.net/cgi-bin/wiki.pl?WikiHomePage#nid1J    (01)

<Prev in Thread] Current Thread [Next in Thread>