Running an industry Machine Learning conference the academic way

Calls for Proposals with a blind reviewing process are more common to academic conferences than industry ones, but there are several important web conferences such as RailsConf that use the same kind of process (we actually use their open source system -- see our previous article). The idea is that proposals are reviewed by members of the program committee who do not know who the authors are, nor which organizations they're from (hence "blind reviewing"). This year we had 2 program chairs who succeeded to Francisco Martin (our '14 and '15 program chair) and who were in charge of turning reviews into decisions of acceptance/rejection: Keiran Thompson and Florian Douetteau. Those decisions are based on average ratings and on "Strong Accepts" given by reviewers. I invited Florian and Keiran to be program co-chairs after consulting with the PAPIs advisory board, and this was before Dataiku became a sponsor of the '16 edition -- it goes without saying that invitations to the committee are unconditional to sponsorship of any kind.