[Air-L] Machine Learning + Image-making

[Edward Summers]

Hi Abi,

If you haven’t run across his work yet Trevor Paglen [1] is doing some really amazing work with machine vision [2,3]. Also, James Bridle is an artist who has been writing about machine learning and culture for some time, and has a new book out [4]. Finally, I just finished Adrian Mackenzie’s book Machine Learners [5] which takes a deep dive into the nuts & bolts of machine learning, and uses critical theory to examine how machine learning is used as a form of knowledge production and a strategy of power. I think some of the ideas he develops in the book could be really useful for the arts/humanities.

Just as an aside, I think one of the challenges for artists working in this area is description. How do you present information about the data and processes that yield the work for a general audience, while also providing enough detail for others who would like to repeat or extend your work? I’m reminded of of Sol LeWitt’s famous instructions for generating art [6], which became works of art themselves [7]. I think effectively communicating these descriptions can lead to greater appreciation. But I could also see an argument for obscuring it, and letting the viewer try to connect the dots themselves.

Good luck with your project, I hope you will share some of your results here!

//Ed

[1] http://paglen.com/
[2] https://www.theartnewspaper.com/feature/trevor-paglen-machine-visions
[3] https://vimeo.com/205149078
[4] https://www.versobooks.com/books/2698-new-dark-age
[5] https://mitpress.mit.edu/books/machine-learners
[6] https://massmoca.org/sol-lewitt/
[7] http://solvingsol.com/solutions/