Illuminating ‘the ugly side of science’: fresh incentives for reporting negative results

New data repositories and alternative journals and workshops offer routes for sharing negative results — which could help to solve the reproducibility crisis and give machine learning a boost.