Critical Machine Learning

Machine learning algorithms are becoming increasingly important elements of the computational processes operating in the contemporary society. Their influence calls for discussions about the social implications, not only by technology experts but also by a broader range of people which are affected by the technological change. This in turn is helped by both accessible technical knowledge and critical thinking, which complement each other. There are a lot of openly available resources that treat either the technicalities of machine learning or more socio-humanistic critiques of the technology. Efforts to mediate the space in-between are also more and more present, but I feel that another such attempt couldn’t hurt—hence this list.

This online resource contains selected readings that can help understand basic concepts of machine learning/AI, as well as its characteristics as a technology situated within social contexts. It is accompanied by this account on The two resources mostly overlap; the account is updated more frequently, whereas this website is intended to be more stable. Although this whole thing is intended to serve as a temporary meta-resource amidst a rapidly changing socio-technological environment, so may become obsolete very quickly. I only hope this is useful in the meantime.

This project was initially conducted as an independent study project for the Interactive Technology and Pedagogy certificate program at the Graduate Center, CUNY where I was advised by the super Maura Smale. Project write-up: link

The list is intended partly in order to serve as a workshop curriculum. My presentation on this topic at the 2017 NYCDH Week: link


The items here were chosen under two criteria: 1) the material should be produced in a way that assumes no prior in-depth technical knowledge, nor the intention to learn it in the future, from the readers; and 2) each material, or some materials combined, should provide an overview on the social implications that this technological trend represents.

Update: Cambridge Analytica

I’ve started gathering some readings relevant to Cambridge Analytica and broader data surveillance in this channel.

Short introductions

Daly, Liza. “AI Literacy: The basics of machine learning,” World Writable, Apr 11, 2017.
At this point there are so many introductions to machine learning. I picked this one because it is short, easy-to-read, and it only does one thing: explaining the confusing terms that are artificial intelligence, machine learning, neural networks, and deep learning.

Artificial Intelligence, Revealed (Video series by Facebook)
Along with the previous post, I think it serves as a good lead into the following three longer reads.

Historical context and recent rise of machine learning

Lewis-Kraus, Guideon. “The great AI awakening.” The New York Times Magazine, Dec 14, 2016.
#translation #nlp #neuralnets #google

Mukherjee, Siddhartha. “A.I. VERSUS M.D.” The New Yorker, April 3, 2017.

A Return to Machine Learning by Kyle McDonald
A “survey of recent developments in machine learning research that intersect with art and culture.”

More context on the role of software code in the technology industry

Tanz, J. “Soon We Won’t Program Computers. We’ll Train them Like Dogs.” Wired, May 17, 2016.

Ford, Paul. “What Is Code?” Bloomberg Businessweek, June 11, 2015.
Not exactly ML-related, but fits well with the Tanz piece above

Huet, Ellen. “The Humans Hiding Behind the Chatbots.” Bloomberg Technology, April 18, 2016.  

Gray, Mary L. and Suri, Siddharth. “The Humans Working Behind the AI Curtain.” Harvard Business Review. January 09, 2017.

A bit more technical introductions, with diagrams and such

Machine Learning is Fun! by Adam Geitgey

A Visual Introduction to Machine Learning by R2D3
Visualizes the decision tree algorithm.

Handwriting recognition:
Back to the Future of Handwriting Recognition by Jack Schaedler
This essay explains a non-ML approach that used hard-defined rules to determine what the user wrote. In contrast, the following two links (note: quite heavy on the browser) show machine learning-based approaches at a similar task:

An Interactive Node-Link Visualization of Convolutional Neural Networks by Adam Harley

Visualizing MNIST: An Exploration of Dimensionality Reduction by Christopher Olah

A Visual and Interactive Guide to the Basics of Neural Networks by J Alammar

Machine learning as black box

Read more on this topic

O’Neil, Cathy. “Welcome to the Black Box.” Jacobin, September 19, 2016

Burrell, Jenna. “How the Machine ‘Thinks’: Understanding Opacity in Machine Learning Algorithms.” Big Data & Society 3, no. 1 (2016): 10.1177/2053951715622512.

FairML: Auditing Black-Box Predictive Models by Julius Adebayo

Knight, Will. “The Dark Secret at the Heart of AI.” MIT Technology Review. April 11, 2017.

Angwin, Julia et al. “BREAKING THE BLACK BOX.” (Series of 4 videos and articles) Propublica. 2016.

Machine learning as reproducer of existing norms and injustice

In what case ML / AI serves to reinforce existing norms, prejudices, injustice, and such? How does that happen and what can we do about it?

Angwin, Julia et al. “Machine Bias.” Propublica, May 23, 2016
related interview:

Stroud, Matt. “Chicago’s predictive policing tool just failed a major test.” The Verge, Aug 19, 2016

Bolukbasi, Tolga, Kai-Wei Chang, James Y. Zou, Venkatesh Saligrama, and Adam T. Kalai. “Man is to computer programmer as woman is to homemaker? debiasing word embeddings.” In Advances in Neural Information Processing Systems, pp. 4349-4357. 2016.

Some tweets I ran into:
“a cool thing to remember is that whenever someone says ‘A.I.’ what they’re really talking about is ‘a computer program someone wrote’”

“tech that helps us CONFORM” (referring to Google Autodraw)

“Machine learning is like money laundering for bias”

Other lists

ML/AI-related things are pouring out in a ridiculous speed, reflecting the rapid growth of the field both as industry and research. Accordingly, people have built lists, daily digests, aggregators and other projects in order to collect and filter information. Someone is probably building a meta-list of these collections right now, if they haven’t done so yet, and that it will not be long until that meta-list itself is outdated. It amuses me how a crucial piece of machine learning is dealing with large data, but I am hardly able to deal with the amount of machine learning-related information.

Anyway, this list points to other collections that I find relevant or complementary to the cml list, that inspired me, that I mean to take a closer look, etc.

Critical Algorithm Studies: a Reading List by Tarleton Gillespie and Nick Seaver

Data & Society’s links page

Propublica’s Machine Bias (series)


The Non-Technical Guide to Machine Learning & Artificial Intelligence by Sam DeBrule
This list gives an industry-oriented overview.

If you want to learn the technical stuff, the following links are probably helpful to you:
Selection of resources to learn Artificial Intelligence / Machine Learning / Statistical Inference / Deep Learning / Reinforcement Learning by Memo Atken

Machine Learning for Artists by Gene Kogan and Francis Tseng

The Programming Historian (especially the Distant Reading section)

Seeing Theory by Daniel Kunin
Not precisely ML-related but nicely visualizes some basic statistical concepts.

A.I. Experiments by Google
Online gallery of AI-based art/design projects, curated by Google.

Cybernetics Conference

AI Now Institute