Carnegie Mellon University professor Tom Mitchell and the Massachusetts Institute of Technology's Erik Brynjolfsson expect machine-learning (ML) computer systems to have a transformative economic impact, and they have outlined 21 criteria for assessing whether a task or occupation is amenable to ML. They say the skills people opt to nurture and the investments businesses make will shape who prospers and who doesn't once ML is incorporated into everyday life.
Mitchell acknowledges ML's effect on a particular job or profession can be difficult to predict because the technology tends to automate or semi-automate individual tasks, while jobs often entail multiple tasks. Mitchell and Brynjolfsson expect ML-amenable tasks to include those for which large datasets are available, such as credit card fraud detection requiring training on many examples. However, they note ML is a not a good option if the user requires a detailed explanation for how a decision came to be made.
From Carnegie Mellon University
View Full Article
Abstracts Copyright © 2017 Information Inc., Bethesda, Maryland, USA
No entries found