RETHINKING ML PAPERS

Over the last decade, the volume of conference submissions in machine learning has broken records across the sciences. According to some estimates, submissions at machine learning conferences account for over 0.5% of all scientific papers published today. While submissions are at an all-time high, there is growing concern in the ML community about where the field is headed. The current pandemic gives researchers a long-awaited opportunity to pause and reflect: what kind of legacy do we want to leave behind? Is it in the number of papers we publish, or the number of citations we accumulate? Or, is it in the clarity and transparency of scientific communication? We believe the answers to these questions will shape the next few years of ML research... How might we improve the scientific paper to create a more sustainable platform for future contributors? Many proposals have emerged, from novelty and rigor (Sculley et al., 2018), to reproducibility (Pineau et al., 2020), explainability..