For instance, OpenFace uses dlib for face detection which was trained on Labeled Faces in the Wild images which in turn used the Viola-Jones face detection that relies on Haar cascade classifier (no longer state-of-the-art) and these images were scraped from internet almost 10 years ago. … Mehr…

Deep Saturation

Vectorization, the deep saturation of the table by algebra, constitutes all relations as movements of transformation, diagonalization, inversion or rotation. Mackenzie 2017:76

Traversing Date

Drawing lines or flat surfaces at various angles and directions is perhaps the main way in which the volume of data is traversed, and a relation between input and output, between predictors and prediction, consolidated as a loci or data … Mehr…


The table or the row-column addressable grid is common to all of [machine learning] datasets. And yet, as we are about to see, machine learning in many ways deals with the collapse or liquidation of tabular datasets. Mackenzie 2017:57 #table … Mehr…

Vectoral Space

The operational power of machine learning locates data practice in an expanding epistemic space. The space derives, I will suggest, from a specific operational diagram that maps data into a vector space. Mackenzie 2017:52

Statistical algorithmic diagrams

In #machinelearning, coding changes from what we might call symbolic logical diagrams to statistical algorithmic diagrams. Mckenzie 2017:46

Coding cultures

…the use [of] programming languages such as Python and R more than specialized commercial statistical and data software packages such as Matlab, SAS or SPSS  is perhaps symptomatic of shifts in computational culture. Coding cultures are crucial to the recent … Mehr…


Materially, code is only one element in the diagram of machine learning. It displays, with greater or lesser degrees of visibility relations between a variety of forces (infrastructures, scientific knowledges, mathematical formalisations, etc.). McKenzie 2017:45 #AI #ML


an immense constellation of documents, software, publications, blog pages, books, spreadsheets, databases, data centre architectures, whiteboard and blackboard drawings, and an inordinate amount of talk and visual media orbit around #machinelearning. McKenzie 2017:44 #AI

Not magic

#Machinelearning is not magic; it cannot get something from nothing. What it does is get more from less. Programming, like all engineering, is a lot of work: we have to build everything from scratch. (Domingos 2012, 81)