Welcome! Please contribute your ideas for what challenges we might aspire to solve, changes in our community that can improve machine learning impact, and examples of machine learning projects that have had tangible impact.
Lacking context for this site? Read the original paper: Machine Learning that Matters (ICML 2012). You can also review the slides.
16 years ago, the same issues were observed
  • I've just learned of an essay written by Foster Provost, Tom Fawcett, Andrea Danyluk, and Patricia Riddle called "On the value of applied research in Machine Learning."  It was written in response to some observed challenges in the lack of (published) ML work that connected to real-world problems.  The essay reads as if it were written yesterday -- making a lot of points identical to, or highly related to, ones in my paper.  What's disturbing is that this essay was posted in 1996.  Why are we still seeing the same problems today, 16 years later?

    Now I'm wondering if some people's suggestion that the ML community split into two communities, along this issue, makes more sense than I originally thought.  If we simply can't agree on ways to reward work that connects theoretical and algorithmic advances with real-world impact, then what other option is there?

    What do you think?  Have we exhausted our options?  Fully explored and discussed them and failed?  Is my paper too little, too late?

  • 1 Comment sorted by
  • Well, I know you started this one on email but I'll start it up here too :)  

    I don't think it makes sense to split.  But I think maybe ICML needs to change or there will at least be the split there.  I think KDD and AAAI have the applications and theory mixture working better.  Perhaps ICML needs to look at how that works and adopt it?  Or perhaps they are happy with how it works and the rest of us just need to take our work elsewhere?  That is what I've chosen but I admire you for trying to change things.


To post or add a comment, please sign in or register.

Tip: click the star icon to bookmark (follow) a discussion. You will receive email notifications of subsequent activity.
If search doesn't work, try putting a + in front of your search term.