My collaborators are blind to my fears.
I am afraid of the data scientists who actively abandon their humanity. An algorithm that blindly optimizes for one objective can discriminate minorities for mortgages rates, spread inflammatory stories, or accidentally reveal information you kept private. Real problems are rarely solved by tackling a single objective, so why do we trust algorithms designed to be narrowly focused? Somewhere we lost the imagination for consequences and being empathetic for others.
I am afraid of data scientists who forget the basics. In the pursuit for fancier and more complex models like deep learning, people have forgot to ask how the data was collected, forgot to articulate their scientific hypothesis before looking at the data, and forgot to relate their analyses to an actionable problem.
I am equally afraid of dance dying as an art form.
Dancers are being priced out of San Francisco despite sharing rooms well into their 30s. At dance performances, a good fraction of the audience members seem to be other philanthropic dancers. In the pursuit of art, did we somehow forget the audience or did we lose to the two dimensional screens that erase our three dimensional sense?
I want to discuss how dance and data science can come together. Dance embraces the infinite possibilities in space yet also how the most mundane gestures can become dance. Data science is constantly looking for potential in different types of data. Can these two worlds come together?
Wayne Tai Lee, Columbia University Department of Statistics