machine learning

machine learning

1. scientific-engineering devices. operational algorithms - "learn from experience". all take shape as computer programs or codes to learn from data

   • data science pattern recognition data mining artificial intelligence
   • across disciplines: from business, health to media and entertainment. ML found in cancer breast research, computer gaming, face recognition systems...

2. from institutionalized to mundane mechanisms recognize gestures, voice, computer games console, SIRI.

   • machine learning in epistemic, economic, institutional settings. redefinition of knowledge, of control, of decisions?
     • Author, Archeology of Operations. Foucault. ML as form of knowledge production and a strategy of power. No exteriority between techniques of knowledge and strategies of power,.
       • ML as classifiers. Couldry and categories. produce power by enacting categories that mark and divide the world in certain ways.
       • Diagrams. archeologically diagrams the data practices of ML with respect to knowledge production, Deleuze. a display of relations between forces which constitute power. Social cultural theory. algorithmic. virtuals that generate variety of actuals. they are hidden.
     • Asymmetries in common knowledge. skip the maths for the beauty of the result. intellectual work using ML in governmental media scientific fields, without observation how changes knowledge production

3. 3 Accumulations. Settings, data, devices

4. positivity of knowing (Foucault)

5. ML viewed as change in how programs, or code that controls computer operations, operate. "With ML, computers writes their own programs, so we don't have to"

   • "control revolution"
     • Amoore: critical enquiry on calculations that will automate future decisions. Some potential futures are "ruled out". Munster: "predictions takes down potential"
   • Author: ML is a convoluted historically specific form of calculation (algorithms) that works to mediate future oriented decisions
     • Author: claim for critical enquiry on automation and calculation notions of ML. Predictive and classificatory calculations with all the investment that attracts (institutions, gov, corporations), rule out and reinforce some futures.
       • Author claim II. algorithm concept. related to changes in media, business, gov, health, science. so: algorithms are powerful. Abstraction.
         • moving at same level of abastraction. Pasquinelli. vs Author: not possible. given accumulation there might not be any single level of abstraction to move in