The NeurIPS Retrospectives Workshop is about reflecting on machine learning research. In addition to publishing meta-analysis papers on the state of the field, this workshop will kick-start the exploration of a new kind of scientific publication, called retrospectives.

What is a retrospective?

A retrospective is written about a single paper, by that paper’s author, and takes the form of an informal blog post. The purpose of a retrospective is to answer the question:

“What should readers of this paper know now, that is not in the original publication?”

The overarching goal of retrospectives is to do better science, increase the openness and accessibility of the machine learning field, and to show that it’s okay to make mistakes. We are accompanying the workshop with the open-source release of a retrospectives platform on GitHub called ML Retrospectives, which will host retrospective submissions going forward after NeurIPS 2019.

How do I submit?

There are two tracks for submissions to the workshop (see: call for papers):

  1. Retrospectives track. This is for submitting retrospectives about your own past papers. They can include new perspectives on the work since publication, a discussion of the strengths and limitations of the work, or acknowledgement of errors in the original manuscript. The goal is to be as open and honest about your past work as possible.
  2. Meta-analysis track. We also encourage the submission of meta-analyses of a small set of related papers (not necessarily your own). This can include observations of recent trends, a discussion of changes to the ‘common knowledge’ in the field, or something else. The goal is to spark discussion about how we do research.

To submit to the Retrospectives track, go to the ML Retrospectives website here, and indicate in your pull request that you want your retrospective to be reviewed for the NeurIPS 2019 workshop. To submit to the Meta-analysis track, go to the OpenReview link here. This OpenReview link can also be used to submit retrospectives if you don’t have a Github account.

Submissions can be from any subfield of machine learning or related fields of interest to the NeurIPS community, including neuroscience and statistics. The main goal of the workshop is to widen what is publishable in ML, and to introduce researchers to more public reflections of their work as part of an ongoing effort to disseminate scientific knowledge more effectively and openly.