R&D is at the heart of Opscidia's activities

Addressing disorder in scholarly communication: Strategies from NISO 2021

Leveraging Open Access publishing to fight fake news

As a result of Sylvain Massip’s participation in the NISOplus 2020 Open Science conference, a prototype has been designed to identify, through the automatic analysis of scientific articles, the consensus associated with a medical question.

The prototype includes a statement such as:

The prototype takes a statement such as “Does the vaccine cause autism?” and constructs three indicators to highlight the scientific consensus behind this statement

BagBERT: BERT-based bagging-stacking for multi-topic classification

The task was to classify articles related to Covid-19 according to their topic. Is the article about predicting the number of cases? Testing a new drug? Or researching the mechanisms of transmission? The system we have proposed performs this complex task automatically and quickly, facilitating the selection of relevant content and saving considerable time in the search for information.

Illustrating the value of Open Access for text mining

In 2018, open access accounted for about 28% of total published scientific articles. While the cost of publication is widely discussed, there is not much work on the value that open access to scholarly publications can actually bring. The aim of this study was therefore to analyze the additional value that full text can bring to text mining tools.