Tutor: Alexander Fefegha
Tutor email: ********@ual.ac.uk
UAL CCI Slack: @AlexFefegha
Session Times: Wednesday 9.30-1.30
Location: CCI 5th Floor Block B
This unit explores the emerging area of machine learning (ML) and its potential impact on culture and society. The unit is a mix of practical tasks introducing ML frameworks such as TensorFlow and seminars that look at emerging practice across the arts and creative industries that employ some level of artificial intelligence. This unit will also explore the centrality of ‘the network’ to computational experience and how machine learning is extending its scope and reach.
From this exploration we will use your new material understanding of machine learning methods and your developing critical framework to question cultural assumptions regarding artificial intelligence and to speculate in writing about emerging computational futures. This primary aim of this unit is to enable you to look past the hype of ‘AI’ and develop your critical framework for thinking about computational technology.
This module will be at the intersection of speculative design, experiential futures, interaction design, critical thinking & creative coding.
My goals for the course for you is:
- To engage in critical creative practice with technology.
- To be philosophers who don’t write down ideas, but instead make objects that embodied them.
- To engage in rapid exploration and experimentation.
- To explore futures and shape alternative futures that is different from the norm.
We will explore things with studio work and seminar-style discussion.
Throughout the module, our tangibles will include:
- mockups
- videos
- prototypes
- design fictions
- interaction designs
On completion of this unit you will be able to:
- Understand how machine learning work in practice (Knowledge, Process)
- Understand artificial Intelligence as a cultural concept (Enquiry)
- Critically discuss computational futures (Enquiry, Communication)
Unit Assessment Summary - This unit is assessed holistically
Assignment Description (2,000 words essay)
-
Using examples, identify and develop an essay containing three case studies that exemplify an argument of how artificial intelligence is culturally constructed or explained.
-
Your case studies should triangulate a key argument that explores how narratives, myths and rhetoric develops around AI and how these are challenged or counteracted. misused or exploited, where deviance enters the use case, or where they are used out of their assumed context.
-
Your case studies may be drawn from commercial, activist, artistic or other fields.
-
The best way to approach your essay is by doing a 400 word introduction (What is machine learning? How do machines learn? AHistorical overview?), 400 words per case study (Work by artists? Work by ML researchers? Work by authors? Work by designers?Work by governments? Work by companies?), 400 words conclusion (Based on all you absorbed and conducted, what would your artistic speculation provocation on machine learning be?)
- To be human.
- There is no right or wrong.
- We are here to learn and have fun.
- Collaboration is everything. People are cool.
- Respect everyone and their difference.
- Give everyone a voice, recognize your privilege and be an ally.
- Challenge each other in nice ways :)
Process Log:
- I would love for you to write your thoughts and learnings somewhere. I use medium but it could be github, wordpress blog, twitter, insta, are.na or what every butters your bread.
- Write about the topics discussed in class, projects that inspire you, and the experimentations you will be doing.
- This is your space to reflect and express yourself. It is your digital sketchbook. Decorate your blog however you choose, use an informal tone, scan your drawings, use memes --- so far as the content is intelligible, organised, and shows critical engagement with content of the module.
Research Presentation Tips (from Irene Fubara-Manuel):
- Contextualise the chosen subject, within the proper historical, political, cultural, and artistic environment.
- Illustrate with videos, images, and text, the key pieces in the subject's portfolio.
- Highlight the tools, processes and motivations of the subject.
- Expand on how or why this topic speaks to you.
- Bring 2-3 questions from your research that we can discuss in the seminar.
- Submit your presentation as a PDF of your slides uploaded onto your GitHub repository.
Core Text:
Bratton, B.H. (2016) The Stack: On Software and Sovereignty. MIT Press.
Engelbart, D. (1962). Augmenting Human Intellect: A Conceptual Framework.
Dourish, P. (2017). The Stuff of Bits: An Essay on the Materialities of Information. MIT Press.
Karparthy, A, Hacker’s guide to Neural Networks
Manovich, L. (2013) Software Takes Command. A&C Black.
Montgomery, E. P, & Woebken, C. (2016). Extrapolation factory operator's manual. New York: Extrapolationfactory.com.
Dunne, A. & Raby, F. (2014). Speculative Everything: design, fiction and social dreaming. MIT Press.
Bleecker, J. (2009). Design fiction: A short essay on design, science, fact and fiction, Near Future Laboratory, Los Angeles, CA,
Bleecker, (2011). Design Fiction: From Props To Prototypes, Negotiating Futures / Design Fictions, Swiss Design Network 2011, Basel.
Kirby, D. (2010). The future is now: Diegetic prototypes and the role of popular films in generating real-world technological development. Social Studies of Science 40 (1), pp. 41-70.
O’Regan, G. (2012). A Brief History of Computing. Springer Science & Business Media.
Zuboff, S. (2018). The age of surveillance capitalism: the fight for the future at the new frontier of power. London, Profile Books.
Finn, E. (2017). What algorithms want: Imagination in the Age of Computing. Cambridge, Massachusetts, MIT Press.
Salter, C., & Pickering, A. (2015). Alien Agency: Experimental Encounters with Art in the Making. Cambridge, Massachusetts, The MIT Press.
Noble, S. U. (2018). Algorithms of oppression: how search engines reinforce racism.
O'Neil, C. (2016). Weapons of math destruction: How big data increases inequality and threatens democracy.
Srnicek, N., & Williams, A. (2015). Inventing the future: Postcapitalism and a world without work.
Graeber, D. (2016). The utopia of rules: On technology, stupidity, and the secret joys of bureaucracy.
Greenfield, A. (2018). Radical technologies: The design of everyday life.
Halpern, O. (2014). Beautiful Data: A History of Vision and Reason since 1945. Durham, NC, Duke University Press.
In Nafus, D. (2016). Quantified: Biosensing technologies in everyday life.
Shadbolt, N., & Hampson, R. (2019). The digital ape: How to live (in peace) with smart machines.
Cheney-Lippold, J. (2019). We are data: Algorithms and the making of our digital selves. New York, N.Y: New York University Press.
Perez, C. C. (2019). Invisible women: Exposing data bias in a world designed for men. London: Vintage Books.
Seminar: Introduction Lesson \\ Intro to Alex Fefegha's practice and his work exploring AI & creativity. We will then read "Man Computer Man-Computer Symbosis" & share our thoughts on it.
Homework:
-
Read: N/A
-
Using the principles thought in class: N/A
Recommended Reading:
Licklider, J.C.R., "Man-Computer Symbiosis", IRE Transactions on Human Factors in Electronics, vol. HFE-1, 4-11, Mar 1960. Eprint.
Studio: Experiential Futures \\ An introduction to Experiential Futures, a brief history on future studies, exploring how futures and design intersect with each other & conducting a ethnographic experiential futures exercise.
Homework:
-
Read: Future-Oriented Design
-
Using the principles thought in class: N/A
Recommended Reading:
Montgomery, E. P, & Woebken, C. (2016). Extrapolation factory operator's manual. New York: Extrapolationfactory.com.
Dunne, A. & Raby, F. (2014). Speculative Everything: design, fiction and social dreaming. MIT Press.
Bleecker, J. (2009). Design fiction: A short essay on design, science, fact and fiction, Near Future Laboratory, Los Angeles, CA,
Bleecker, (2011). Design Fiction: From Props To Prototypes, Negotiating Futures / Design Fictions, Swiss Design Network 2011, Basel.
Kirby, D. (2010). The future is now: Diegetic prototypes and the role of popular films in generating real-world technological development. Social Studies of Science 40 (1), pp. 41-70.
Studio: The Anatomy of An AI System \\ An introduction to machine learning + creativity and the tools/frameworks used to make cool stuff! (ML5.js)
Homework:
-
Watch: Body, Movement, Language
-
Read: Real-time Human Pose Estimation in the Browser with TensorFlow.js
-
Make: Taking the work we did in class using the Ml5 library for PoseNet. I would like for you to build a clown face. Here's my example from class
Recommended Reading:
Crawford, K. and Joler, V. (2018). Anatomy of an AI System.
Hertzmann, A. (2018). Can Computers Create Art? Arts 7, 18
Agüera y Arcas, B. (2017). Art in the Age of Machine Intelligence. Arts, 6, 18
Resources:
Pose Estimation (Tensorflow.js)
Maya Man (Creative Technologist)
Seminar/Studio: How do we as creatives explore with a future with AI? \\ An introduction to Future Crafting, A Speculative Method for an Imaginative AI by Betti Marenko and exploring ImageNet, and creating our own image classifier with ML5 & P5.
Homework:
With the image classifier that you created in class, try it out on ten different objects around you. Use this list of 1000 ImageNet Classes to guide you too.
- Does the image classifier give a high confidence score?
- Did it succeed in recognizing an object?
- Think about the different things that could or may affect the image classification?
Recommended Reading:
Marenko, B. (2018) FutureCrafting. A Speculative Method for an Imaginative AI. In: AAAI Spring Symposium Series. Technical Report SS-18. Association for the Advancement of Artificial Intelligence, Palo Alto, California, pp. 419-422.
Rohilla, H (2018) Designing For Future: Speculative Design Research For A Data Obsessed Society.
Resources:
Image Classification using Deep Neural Networks — A beginner friendly approach using TensorFlow
Studio: Invisible Mask \\ An introduction to Alex's project exhibited at MozFest2019 which was a speculative provocation exploring the attack on human agency and autonomy by facial recognition technologies and the introduction to retraining a machine learning model.
Homework:
Try to finish the feature extraction with image classification example that Alex was teaching in lesson using Visual Studio Code.
The p5.js sketch is here
Recommended Reading:
de Vries, P. and Schinkel, W. (2019) ‘Algorithmic anxiety: Masks and camouflage in artistic imaginaries of facial recognition algorithms’, Big Data & Society.
Blas Z (2014) Informatic opacity. The Journal of Aesthetics & Protest.
Blas Z (2016) Infomatic opacity. In: The Black Chamber: Surveillance, Paranoia, Invisibility & the Internet. Brescia: Link Editions & Ljubljana. Aksioma, pp. 40–51.
Resources:
How to build a Teachable Machine with TensorFlow.js
Studio: Recap \\ An introduction to the essay for this module, a recap on what has been taught so far & play with PoseNet again!
Homework:
Make a full body experience emoji creator with PoseNet!
- See what you could do with p5.js capabilities (like making shapes of the body).
- See what you could also do with Teachable Machine if you want to push the boundaries!!!
Recommended Reading:
Schmitt, P. (2019). Humans of AI
Hebron, P. (2016). Machine Learning for Designers
Resources:
A visual introduction to machine learning
Move Mirror: An AI Experiment with Pose Estimation in the Browser using TensorFlow.js
ml5: Friendly Open Source Machine Learning Library for the Web
CCI's very own Rebecca Fiebrink talking about Machine Learning for Human Creative Practice
Interactive Seminar: It's all in the data! \\ The class will be an introduction by Alex to creating data sets to train a machine learning model. Alice Stewart will be sharing about her latest creative machine learning project.
Homework:
Think about how machine learning can assist you in your project for your Creative Practice: Visual Coding and Physical Computing Class. This should be put together for a 3 minute presentation next week!
- Think about what are the data sets you need?
- Do you need to create them from the blank canvas or you could source data sets from else where?
Recommended Reading:
N/A
Resources:
What is the Difference Between Test and Validation Datasets?
Tabular Data - Working With Data & APIs in JavaScript
How to Build A Data Set For Your Machine Learning Project
How I'm fighting bias in algorithms by Joy Buolamwini
How to keep human bias out of AI by Kriti Sharma
Racial Bias and Gender Bias Examples in AI systems
Studio/Lecture: Neural Networks \\ Introduction to neural networks, how to construct your own DIY neural network and train a classification model.
Homework:
-
Watch the series on Color Classifier by the Coding Train on Youtube.
-
Train a model on a color classifier data set in p5 + ml5.js, not Tensorflow.js!
-
I have left you an example of my code.
Recommended Reading:
Boden, M. A. (2009). Computer Models of Creativity. AI Magazine, 30(3), p.23.
Resources:
Nature of Code - Chapter 10. Neural Networks
Using neural nets to recognize handwritten digits
Perceptron — Deep Learning Basics
Neural Networks by Google Developers
Activation Functions in Neural Networks
Neural Networks - Intelligence and Learning by The Coding Train
The Danger of AI is weirder than you think by Janelle Shane
Studio/Lecture: CNNs? \\ Introductions to Convolutional Neural Networks and create your doodle classifier using Google's Quick Draw datasets.
Homework:
Think about what you want to explore in your essay. I would like for you to create a slide deck consisting of:
- a research topic regarding AI
- ideas about how you might intro
- ideas of case studies
- ideas of how you might conclude
- any questions you have
Please do send to me on slack for the 4th of Dec. It is not a presentation!
Recommended Reading:
Gebru, T., Morgenstern, J., Vecchione, B., Vaughan, J. W., Wallach, H., Hal Daumé, I., et al. (2018). Datasheets for datasets.
Resources:
Convolutional Neural Networks - Deep Learning Book
An intuitive guide to Convolutional Neural Networks
Exploring and Visualizing an Open Global Dataset
A visual and intuitive understanding of deep learning (video)
On Missing Data Sets by Mimi Onuoha
Machine learning & art - Google I/O 2016 (video)
Atlanta Asks Google Whether It Targeted Black Homeless People
AI ‘EMOTION RECOGNITION’ CAN’T BE TRUSTED
Studio/Lecture: RNNs? \\ Introduction to Recurrent Neural Networks and generate music using a neural network.
Homework:
Rest over the holidays.
Maybe do a sketch a day.
Recommended Reading:
Amato, Giuseppe & Behrmann, Malte & Bimbot, Frédéric & Caramiaux, Baptiste & Falchi, Fabrizio & Garcia, Ander &Geurts, Joost & Gibert, Jaume & Gravier, Guillaume & Holken, Hadmut & Koenitz, Hartmut & Lefebvre, Sylvain & Liutkus, Antoine & Lotte, Fabien & Perkis, Andrew & Redondo, Rafael & Turrin, Enrico & Viéville, Thierry & Vincent, Emmanuel. (2019). AI in the media and creative industries.
Asay, Clark D., Artificial Stupidity (June 4, 2019). 61 William & Mary Law Review (2020, Forthcoming). Available at SSRN: https://ssrn.com/abstract=3399170
Resources:
Illustrated Guide to Recurrent Neural Networks
Rohan & Lenny #3: Recurrent Neural Networks & LSTMs
Neural Nets for Generating Music
Computer evolves to generate baroque music!
Recurrent Neural Network Tutorial for Artists
Four Experiments in Handwriting with a Neural Network
Connecting with Music Through Magenta.js