CONCORD: Numerical Claims Extracted from the COVID-19 Literature using a Weak Supervision Approach

The COVID-19 Numerical Claims Open Research Dataset (CONCORD) is a comprehensive, open-source dataset that extracts numerical claims from academic papers on COVID-19 research. To extract numerical claims, a weak-supervision based model is employed, leveraging its white-box, explainable nature and advantages over transformer-based models in terms of computational and manual annotation costs. Labelling functions are used to programmatically generate labels, incorporating techniques like pattern matching, external knowledge bases, phrase matching, and third-party models. An aggregator function reconciles overlapping or contradictory labels. The weak-supervision model is evaluated against established baselines and transformer based models, achieving a weighted F1-score of 0.932 and micro F1-score of 0.930 in extracting numerical claims. In comparison to baseline models, the weak-supervision model outperforms them, although it is noted that transformer-based models demonstrate comparable performance. CONCORD, comprising around 200,000 numerical claims extracted from over 57,000 COVID-19 research articles, is valuable for knowledge discovery and understanding the chronological developments in various research areas. The study highlights the potential of weak-supervision models for Argument(ation) Mining in the biomedical domain, showcasing their efficacy and utility. Overall, CONCORD and the weak-supervision methodology provide a valuable resource for researchers and contribute to advancements in COVID-19 research, while also underscoring the potential of weak-supervision models in the biomedical field.