This year we will focus class 5 purely on reviewing the openFrameworks assignment. The last two years we ran a session on the use of AI in Creative Coding, and given that I know that many of you are interested in this topic, I am leaving this content here for who is interested. If time allows, I will run a short demo session on this topic instead.
Background literature
Let an ultraintelligent machine be defined as a machine that can far surpass all the intellectual activities of any man however clever. Since the design of machines is one of these intellectual activities, an ultraintelligent machine could design even better machines; there would then unquestionably be an 'intelligence explosion,' and the intelligence of man would be left far behind. Thus the first ultraintelligent machine is the last invention that man need ever make, provided that the machine is docile enough to tell us how to keep it under control
I.J. Good, "Speculations Concerning the First Ultraintelligent Machine", Advances in Computers, vol. 6, 1965
It's tempting to dismiss the notion of highly intelligent machines as mere science fiction. But this would be a mistake, and potentially our worst mistake ever.
Stephen Hawking, Huffington Post
There needs to be a lot more work on AI safety
Tesla Motors and SpaceX CEO Elon Musk said during a recent Q&A on Reddit.
We may be on the verge of creating a new life form, one that could mark not only an evolutionary breakthrough, but a potential threat to our survival as a species
Jeff Goodell. Inside the Artificial Intelligence Revolution: A Special Report, Pt. 1. Rolling Stone, Feb 29 2016.
Like other forms of violence, algorithmic violence stretches to encompass everything from micro occurrences to life-altering realities.
Mimi Onuoha. Notes on Algorithmic Violence. GitHub.
I don’t work on preventing AI from turning evil for the same reason that I don’t work on combating overpopulation on the planet Mars
Andew Ng, a leading machine learning researcher at Baidu and Stanford University, quoted in Alexis C. Madrigal. The case against killer robots, from a guy actually working on artificial intelligence. Fusion, Feb 27, 2015.
AI has been overestimated again and again, in the 1960s, in the 1980s, and I believe again now, but its prospects for the long term are also probably being underestimated. The question is: How long is the long term?
Rodney Brooks. The Seven Deadly Sins of AI Predictions. MIT Technology Review, 2017.
The irony is that compared with human intelligence, A.I. is actually the more transparent of intelligences.
Vijay Pande. Artificial Intelligence 'Black Box' is nothing to fear. New York Times (Online), Jan 25, 2018.
AI Winter, Wikipedia
The best way to predict the future is to change it.
George Church, leading scientist in genomics and genetic engineering in National Geographic,
There is plenty of room at the bottom
Richard Feynman, in a lecture about miniaturization and the nano scale given at an American Physical Society meeting at Caltech on December 29, 1959. Transcript published at Caltech Engineering and Science, Volume 23:5, February 1960, pp 22-36
For more see the Futures section in Further Reading & Viewing
Lab
Core concepts and inspiration examples: Teachable Machine & AI Experiments
Train your own machine learning classifier with the Teachable Machine and your favorite set of objects to be recognized
Explore some of the other AI Experiments with Google. Which ones do you like purely from a conceptual points of view? Do you have an intuition how these examples may have been built?
Let's get a bit deeper into internals. Have a look at some of the ml5js and TensorFlow.js examples, or any of Gene Kogan's Machine Learning for Artists examples. What are some of the different kind of AI capabilities that we can leverage?
We'll do a quick plenary review of things you have discovered.
Let's build our own models to map sensory information to action: Wekinator
Download and install Wekinator
Peter will do a quick demo
Read the instructions and try out some combinations of input and output in the walkthrough. Do you understand what is happening?
Create your own links between input and output
Check out these Wekinator examples (by Andreas Refsgaard) and/or download the Quick Start Pack
Combine and hack!
If you need a bit more guided examples, see the videos
Alternatively go back to the examples above to experiment and hack.
Finally we will discuss the expo (theme, space, internal page).
And are you into Barbies? Then join the AI AI Barbie Hackathon, on the role and impact of AI in raising kids and parent-kid-toy relationships, on June 6 in De Waag.
Full resource list
ml5js - Simple Javascript library for machine learning inspired by Processing
Presentation by Andreas Refgaard
TensorFlow.js, Javascript library for deep learning (Teachable Machine, Webcam controller, Piano RNN examples are interesting)
Machine Learning for Artists (ml4a), great community site by Gene Kogan
AI and Machine Learning add ons in OpenFrameworks, ml4a-ofx OpenFrameworks examples
Magenta, make music and art using machine learning
Runway, Processing examples, download instructions and link (ask Peter for access)
Google MediaPipe, example of integrating MediaPipe with openFrameworks
General data mining and machine learning toolkits, libraries and resources
Weka - machine learning toolkit (choice of GUIs or scripting/java programming), engine behind Wekinator
sckikit-learn, machine learning in Python
Deep learning
Machine Learning made easy, some older courses for high school students (course 1, course 2, 2005)
KDnuggets.com - general data mining and machine learning community
NeuroHive.io: blog with some of the latest research, data, tools in deep learning