Overview
ASReview LAB is a premier open-source machine learning tool designed to optimize the screening process of systematic reviews. By implementing the 'human-in-the-loop' active learning paradigm, the software significantly reduces the time researchers spend screening irrelevant records by dynamically re-ranking the most relevant papers to the top of the queue. As of 2026, ASReview maintains a dominant position in the research community due to its commitment to transparency and reproducibility. The technical architecture is highly modular, allowing users to select or develop custom components for feature extraction (e.g., TF-IDF, Doc2Vec, BERT), classification models (e.g., Naive Bayes, SVM, Random Forest, Neural Networks), and query strategies. This flexibility, combined with its local-first privacy model where no data leaves the user's machine, makes it the gold standard for sensitive medical, legal, and social science research. The platform supports a wide range of file formats and provides comprehensive visualization tools to estimate when screening can be safely stopped based on recall curves.