User:Manetta/graduation-proposals/proposal-0.0
graduation proposal -0.0
current areas / questions of research
- The presence of machines nowadays, and the difficulties that arise when positioning them into a human centralized world.
- The naturalization of machines, and the invisibility of their interfaces.
- The fact that the architecture of a tool is by default disguised, which is a motivation to open them up and reveal their constructions.
- Metaphors are tools that enable us to speak about abstract elements of the internet. How are they contructed, and where do they refer to?
- Linguistic systems are built to let the computer understand human language. In computer science this is called 'Natural Language Processing' (NLP). Motives to make such a language systemization are paradoxical, as the act of making a model is obviously deductive. Which aims counter-balance this deduction? In other words: what purposes are strived for that justify making a simplified version of reality?
- Significance is mainly constructed through such systems and structures. Signification isn't derived from the objects themselves, but rather from the system they are put in. (Examples: ngrams, valued dictionary files, JSGF files, machine-learning algorithms, newest = deep learning)
- creating an universal language
- classification & taxonomies/ontologies
- problem of an one-sided fixed truth that represents 'common sense'
- naturalization & invisibility of technology as ideology
how to build on them ?
- look closer at how text-mining processes/results are relating to common sense (this summer --> Relearn + Relearn etherpad notes)
- form: making an online 'platform' to start collecting and writing about categorization & systemization (elements of grey zone semantic? turning text in numbers? heteromation? semantic models? )
- describe what to understand as 'machines' --> look at Latour way of approaching artifacts in a sociological way
(im)possible project
My (im)possible project would be a start of an archive for elements that touch the field of semantics. By focussing on structures rather then on objects, i will start collecting (non)fictive examples of categorizations, systemizations and ontologies for example used in fields of Computer Vision (CV?) and Natural Language Processing (NLP).
Making small prototypes based on these examples, would be a good way to shine a light on them. To do so, i will start with selecting a set of 'neutral' objects that will function as placeholders throughout the project.
In computer science, the word 'ontology' is easily mixed up with a term that would fit better to its task: 'taxonomy'. As 'ontology' is a term that comes from a philosophical tradition, it is easy to misunderstand the question of "what is where" (taxonomy) with the question of "what is" (ontology). It is interesting to see how in the field of computation, these terms got mixed up.
(what would be the question/problem to work on?)