Black Boxes Are Not Required

Published: June 5, 2020, 7:59 p.m.

b'

Deep neural networks are undeniably effective. They rely on such a high number of parameters, that they are appropriately described as \\u201cblack boxes\\u201d.

While black boxes lack desirably properties like interpretability and explainability, in some cases, their accuracy makes them incredibly useful.

But does achiving \\u201cusefulness\\u201d require a black box? Can we be sure an equally valid but simpler solution does not exist?

Cynthia Rudin\\xa0helps us answer that question. We discuss her recent paper with co-author Joanna Radin titled (spoiler warning)\\u2026

Why Are We Using Black Box Models in AI When We Don\\u2019t Need To? A Lesson From An Explainable AI Competition




'