Enzymind Labs

misc on multimedia database indexing

=-=-=-added 2006-07-28 psr]

=abstract hilbert spaces

-do attracors live as discrete entities in hilbert space or is there a way of super imposing hilber space with dynamic system phase space (or are phase spaces a subset OF hilbert spaces?)

-use of nonlinear equations as dimentions in hilbert space to model biological systems.

this is kind of what i have in mind in terms of visualization and data exploration:

Hilbert himself described this space by pointing to a glass of beer and explaining that it exists in this space at a place where the parameters describing its shape, size, materials and content intersect. He explained that if you move this object inside the space to where the content parameter is water instead of beer, you would have a glass of water. If this object is moved to a parameter where the material is pewter instead of glass, you'd have a pewter mug full of water. He further explained how by moving this object around in this parameter space you could get it to change into a beer mat, a chair, a table, etc.

The importance of this mental model is that it envisages a space where every possible thing you can ever think of is in there and every one is located at a point where their parameters meet. More importantly, not only does this space contain all the objects you know about, it also contains all the objects you don't know about. [editorial note: this is what blows my mind, because all the things I don't know about are the things I am interested in :)] Even more bizarrely, it also contains objects that are not in existence today but could be in existence in the future.

For example, Google may present you with lists of rules and strategies to cover various kinds of competitive situations. But, the search space is limited to only those lists of rules and strategies that appear on the web sites that have been indexed.

An abstract Hilbert Space will contain all lists of rules and strategies - some of which may not have been published, or, may not have even been discovered yet.

koestler model of creativity. 'bisociation' vs association in creative process, commonality within a matrix domain, bisociation: interesection of two separate domain matrices with each other.

use of interactive visualization tools to model this -image planes in volume representing knowlege domain matrices -maybe as simple as a matrix of images. Ability to have multiple interacting domain matricies, with parallel representations in differnt function spaces; ability to scan spaces to gether and watch planar surfaces or plot the output of the interesection of the different space representations in a separate graph/window.

use of hilbert or other function spaces as place for making connections; transforms; linear transforms; nonlinear manipulations (a la apophysis) ...

=-=-=-added 2006-07-27 psr

IBM marvel

Just thinking about a workflow for how to attack indexing.

Seems like since there is a MySQL data base of images (somewhere) I should be able to parse the data base.

So first thing I need to figure out is how to parse the Database.

Next thing is what form of image should the indexing be done on -mostllikely a thumbnail.

And as was suggested using a most fundamental character such as contrast, spatial frequency, or color as an indexer is probably the best way to start.

Also I need to think through exactly what I intend to acomplish with the image indexing and how to relate it to emergence, creativity, and complex systems...

Any comments or thoughs welcome.