Take care of me

Take care of me, a fictional text written by a machine learning algorithm and a colonial soldier, python script, 9’12”, 2019

Take care of me is a fictional text written by “the most adaptive learning algorithm for machine intelligence”, a neural network (called NN), together with the artist’s uncle Pietro Salvadori (called PS), soldier in the colonial war of the fascist regime in Ethiopia (1935-1936).
The text, displayed one line at a time by a Python script, tells about what the two authors did to become themselves, describing their inner states, purposes and training process. Moreover while the main text runs, we read a “meta-text” shown in parallel windows, where the two writers introduce themselves and make comments on their own words.
NN brings us to follow the key concepts and functions of the statistical inferential model commonly called Neural Network, from an internal perspective, observing the training process and the main mechanisms that define it; from the learning algorithms to the main issues associated to the fitting and training processes.
The soldier Pietro Salvadori speaks about the colonial war, evoking empty spaces and exotic landscapes, where plants, animals and humans were wonderful and stranger at the same time. He tells us about how he became himself in this elsewhere, episodes of this and other wars are evoked through fragments that often mix and overlap until the last image, the suicide, committed during the sixties, by hanging himself in the attic where he used to live.
The two voices that make up the text are constantly mixed and it is often difficult to say who is speaking, the concepts and words are interchangeable, a continuous overlapping of the two flows of consciousness, images and memories. This mechanism creates a shift, a displacement that is the key of the work: the main idea  is to use a fictional text as a mathematical model, both being powerful ways of knowing.

Here the whole text.

Thanks to Eleonora Farina and Luca Sguanci.

Bibliography
Michael A. Nielsen, Neural Networks and Deep Learning, Determination Press, 2015
Stanford University Course CS231n: Convolutional Neural Networks for Visual Recognition, Lectures and Notes, Spring 2018, http://cs231n.stanford.edu/index.html
Matteo Pasquinelli, Abnormal Encephalization in the Age of Machine Learning, e-flux, n. 75, September 2016
Matteo Pasquinelli. Machines that Morph Logic: Neural Networks and the Distorted Automation of Intelligence as Statistical Inference. Glass Bead, 1: “Logic Gate: The Politics of the Artifactual Mind”, 2017. Web: www.glass-bead.org/article/960
Ennio Flaiano, Tempo di Uccidere (Time of Kill), Italian Edition, BUR 2000
Nicola Labanca, Constructing Mussolini’s New Man in Africa? Italian Memories of the Fascist War on Ethiopia, (Italian edition), Il Mulino, 2005
Antonio Orlandi Contucci, Passato d’africa, (Italian only), Rubettino 2011
Alan Turing, Lecture to the London Mathematical Society on 20 February 1947, online at https://www.vordenker.de/downloads/turing-vorlesung.pdf
Gilbert Simondon, On the mode of existence of Technical object, University of Minnesota Press, 2017