What Sandwiches, Categories, and Ethics Have to Do with Machine Learning
Is a burrito a sandwich? According to Bob Kasenchak, director of business development, Access Innovations, Inc., USA, this innocuous-seeming question is a great way of soliciting opinions, and thinking about categories and the power of naming things and its consequences.
In a highly entertaining presentation at Taxonomy Boot Camp 2018, Kasenchak said people make inferences based on semantics (as in a machine learning environment), and therefore putting things into categories becomes an important ethical gesture. As taxonomists name things, make categories, and name categories and put things into categories, he said, they are therefore making ethical decisions as part of their job description.
In a machine learning context, taxonomists must be careful with categories and naming not only because it’s ethical—but also because machine learning is a garbage in, garbage out operation.
The bottom line is that if you are going into, or would like to think about, machine learning in the future as part of your taxonomy program, the implications the structure of your taxonomy are of enormous consequence.
“Machine learning is not a silver bullet,” he emphasized. “It uses data as input and applies it to a model. If your model is poorly constructed, you will get garbage out.”
Many presentations have already been made available online at www.taxonomybootcamp.com/2018/Presentations.aspx and others will become available after the presentations are given.
KM World 2019 will be held November 5-7, 2019 at the JW Marriott in Washington, DC, with pre-conference workshops on November 4.