Shredded STASI file. Aim to get the text (information) from the shredded documents. Uses texture, colour and content details: typscipt, handwritten or stamp print to group shreds together (or rather, the scanned versions of the shreds.
Tuesday, 30 October 2007
Shreds of evidence
Case-based reasoning
AI systems that use a 'case-base', a database of known problem-solution sets, and generate adaptation knowledge to allow solutions to be deduced for new problems.
Anything on AI is bound to have a big information-theoretic content. In this example, you could look at the information content of the case-base, and explore what's going on when generating adaptation knowledge, among other things.
Sunday, 28 October 2007
Identity
This set me thinking about Michael Frayn's exploration of 'self'.
...he goes on to describe the marks and signs people used to identify themselves to others: authenticating seals, coats of arms, trademarks, insignia.
What's the role of information in authentication?
Authorities relied, then, primarily on words to identify people in official documents.
We can of course readily talk about the information content of words
Scars, birthmarks and moles become central to recognition
It is not too difficult to imagine working out the information needed to describe the presence of a scar, birthmark or mole.
Skin colour was also characterised
Again, we can easily come up with a value for the information needed to express skin colour.
...objects of identification issued by authorities for people to wear or carry
ID cards, which today would be digital documents.