Monday, January 12, 11am - 12:30pm

Applied researchers are increasingly using tools and techniques from machine learning and artificial intelligence in their research. While these methods can be extraordinarily powerful, they also pose distinctive challenges for researchers committed to the norms and practices of open and reproducible research. For example, the models that underpin AI research tools tend to be opaque (with respect to both their training data and the algorithms used to “learn” from this data), poorly documented, and yield probabilistic outputs that can be difficult to reproduce. This Workshop provides an introductory conceptual overview of some of the ways in which AI tools could pose challenges for open and reproducible research, and explores ways to address some of these challenges. All are welcome; the Workshop does not presuppose any prior knowledge of reproducible research or artificial intelligence.