Shifts Challenge: Robustness and Uncertainty under Real-World Distributional Shift
Coming summer 2021
We invite researchers from all over the world to participate in our NeurIPS 2021 Shifts Challenge on robustness and uncertainty under real-world distributional shift. The challenge raises awareness of distributional shift in real-world data. The aim is to develop distributionally robust models, which explicitly measure uncertainty in their predictions.
Contestants can take part in three separate tracks, for which we provide datasets in three modalities: weather prediction, machine translation and vehicle motion prediction. Each track represents unique challenges, with top solutions acknowledged and rewarded.