The Scithon™ evaluation framework involves a systematic and cross-disciplinary tool use comparison. During a live event participants are asked to solve a research challenge, with and without the Iris.ai tools. Submitted results are evaluated by domain experts. Through a number of Scithons™ there has consistently been a clear positive relationship between jury percentile score and the percentage of papers found with Iris.ai. Learn more about the framework here.

Word importance-based similarity of documents metric (WISDM), is a fast and scalable method for document similarity computation for analysis of scientific documents. It is based on recent advancements in the area of word embeddings. The evaluation results demonstrate that the WISDM metric achieves significant performance speed-up in comparison to state-of-the-art metrics with a very marginal drop in precision. See the paper here.