Hi organisers,
I just came across this competition. I am a regular participant of Kaggle and other Machine Learning competitions. But I read about this challenge and it looked totally different from the traditional Machine Learning competitions. I want to know whether I can join now and start working in this competition ? Also I a R coder so is there options available to accept R codes ? Sorry if this question sounds very basic because this competition looks pretty different to me.
Thanks
Created by Thanish Batcha thanish > I can join now
Yes, absolutely. The competitive phase has not yet begun. The challenge wiki contains all the instructions and has a link to the September webinar which you can view.
> is there options available to accept R codes ?
Yes, the submission is a Docker container and you can build on top of a base image like one of these:
https://hub.docker.com/_/r-base/
https://hub.docker.com/r/rocker/rstudio/
You *do* have to install on your machine and learn Docker to get started.
Good luck! Welcome to the DM DREAM Challenge Thanish!
> But I read about this challenge and it looked totally different from the traditional Machine Learning competitions.
Yes, the approach taken by this Challenge where the training set is not directly visible to the participants is different from most of the other competitions. This is an additional *challenge* for the Challenge in terms of organization, infrastructure and development of inference methods. However we expect this Challenge to pave the way for future competitions that will make use of datasets that can not be made public (e.g. medical information, data from a company that want to crowdsource a problem, etc.).
> I want to know whether I can join now and start working in this competition ?
Yes, participants can join the competition anytime before the end of the Competitive Phase. Actually the official competition is starting in a few weeks. We are currently in a phase called the Open Phase where participants have untracked quota time to develop a [Docker container](https://www.docker.com/) that embed their inference method and make sure that their method can read the dataset and run on the Challenge Cloud. To avoid participants training their model with unlimited time quota during the Open Phase, the current dataset installed on the Cloud has been generated by duplicating subjects from the [Pilot Data](https://www.synapse.org/#!Synapse:syn6174174), which include 500 mammography images that we have been allowed to release publicly. This _substitute_ dataset is similar in size and format to the training set that will be used during the competitive phase.
> Also I a R coder so is there options available to accept R codes ?
A Docker container is basically a Virtual Machine inside which you can install all the programs and dependencies that you need (R included). For more detailed information on how to build a Docker that include your model and submit it, pease refer to [this page](https://www.synapse.org/#!Synapse:syn4224222/wiki/401759). Please feel free to use this Discussion Forum if you have any questions.
Thanks!
Thomas