Machine Learning (ALA Midwinter 2019)

I gave a talk at the American Library Association 2019 Midwinter conference entitled “Machine Learning: How Does It Work?!?”

Slides: .key (original version), .pdf

Bibliography

If you’re here because I encouraged you to check out my talks page to educate yourself on ML and its discontents, thanks! I extracted the bibliography from the slides for your convenience.

Library AI Examples

• HAMLET
• Blog with background info on HAMLET
• Temple Libraries presentation at LITA Forum 2017 on their ML work — scroll to “Redefining Discovery”
• National Library of Finland automated classification project
• URI AI lab

In Which Hilarity Ensues

• Correlation is not causation
• Numerous examples of AI systems doing exactly what we said (but not what we meant)
• Janelle Shane’s blog, e.g. “When Algorithms Surprise Us”
• xkcd explains machine learning
• Let’s make dessert from a machine-generated recipe

Critical Awareness & Social Impact

• Matthew Reidsma, “Algorithmic Bias in Library Discovery Systems” (if you read only one, make it this)
• Data for Black Lives: @data4blacklives, http://d4bl.org/ — D4BL runs an outstanding conference whose livestreams you can find on YouTube.
• Pro Publica Machine Bias series
• Montana State University IMLS-funded work on Algorithmic Awareness — This will produce an undergraduate course at MSU as well as training materials for librarians at other institutions.
• Syllabus for Dorothea Salo’s “Code and Power” course at UW-Madison
• Meredith Broussard, Artificial Unintelligence: How Computers Misunderstand the World
• Virginia Eubanks, Automating Inequality: How High-Tech Tools Profile, Police, and Punish the Poor — you can also see video of her presenting on the book’s topics
• Safiya Noble, Algorithms of Oppression: How Search Engines Reinforce Racism
• Cathy O’Neil, Weapons of Math Destruction:  How Big Data Increases Inequality and Threatens Democracy

Learn How To Build ML Systems

• Coursera Machine Learning class — This is how I learned enough to build HAMLET.
• deeplearning.ai — next steps beyond Coursera.
• Most papers you might want for going into depth are on the arXiv.
• gensim, the python library behind HAMLET