Tuesday 21 November 2017

The titans of AI are getting their work double-checked by students by @davegershgorn via @qz

In an effort keep AI research a science, Joelle Pineau, an associate professor at McGill University and head of Facebook's AI research lab in Montreal, is pushing back against unreproducible AI research. Her challenge, coordinated with five other universities, is to reproduce the work in submitted papers. Students are tasked with reproducing research from some of the world's top AI labs—from universities to companies like Google, DeepMind, Facebook, Microsoft, and Amazon.

I find it strange that not only is it almost impossible to reproduce the research, but that they publish papers with deliberate missing information.  When learning R and how to publish a paper it was made clear to me that I must publish my code so that others can reproduce my results.

No comments:

Post a Comment

Note: only a member of this blog may post a comment.