Implementing AI Vertical use cases ‐ Scenario 1
Meta data
Disclaimer: The present document has been produced and adopted by the bodies identified above as authors. This task has been carried out exclusively by the authors in the context of a contract between the European Food Safety Authority and the authors, awarded following a tender procedure. The present document is published complying with the transparency principle to which the Authority is subject. It may not be considered as an output adopted by the Authority. The European Food Safety Authority reserves its rights, view and position as regards the issues addressed and the conclusions reached in the present document, without prejudice to the rights of the authors.
Abstract
This report presents the activities and outcomes related to the five use cases (UCs) identified by EFSA within the Specific Contract No 1 implementing the Framework contract No OC/EFSA/AMU/2021/03. UC1 aims to assess the usability and effectiveness of text summarisation tools applied to the context of Public consultations (PCs) with the objective to use AI‐based automatic text summarisation (ATS) techniques to generate a summary for each attachment to the comments received for a PC. UC2 aims at automating the keywords identification step within the systematic review (SR) process, to explore the current capabilities of the existing tools to enhance the keywords identification process to facilitate the subsequent retrieval of potentially relevant studies with the help of machine learning (ML) and artificial intelligence (AI). UC3 assesses the capabilities of deduplication within DistillerSR with the objective to evaluate the effectiveness of different user‐defined deduplication settings available to DistillerSR users to correctly identify and remove duplicates while highlighting the limitations of the tool. The objective of UC4 is to develop a tailored manual for the use of Distiller SR within EFSA's SR framework to facilitate the relevance screening step, replacing human‐based relevance screening of titles and abstract with an AI‐assisted process. UC5 was designed to exploit the capabilities of the DistillerSR Artificial Intelligence SYstem (DAISY) in DistillerSR to build and use classifiers for semi‐automated AI‐assisted characterisation or classification of studies within the SR process. Analogous to UC4, also the UC5 aimed at creating a manual helping the users to benefit from this recently introduced functionality within DistillerSR.