ASReview

ASReview uses active learning to prioritise the records most likely to be relevant during title and abstract screening. Reviewers label records as they work, and the model continually updates the order in which remaining material is presented, helping reduce repetitive screening in large evidence reviews.

The project is coordinated from Utrecht University and documents its methods as open-source research software. It supports transparent review workflows and simulation for testing strategies, but it does not replace reviewer judgement: people still make every inclusion and exclusion decision.