Natural Language Processing supporting Active Learning

The main objective of the eSTEM project is to increase the capacity of lecturers and educators to develop high-quality online training content by increasing their awareness of 4 pedagogical methods (inquiry-based, dataset-based, problem-based, scenario-based) and by providing some “tricks” to follow them easily and quickly.

The “tricks” designed by the partnership are 4 toolboxes, one for each pedagogical method, comprising guidelines, examples of application and, in particular, a dedicated IT tool called SuperFastLearning (SFL) machine. As their names suggest, SFL machines are designed for speeding up the development of STEM learning content, by supporting the analysis of documentation used by lecturers as a technical and scientific source.

SFL machines adopt Natural Language Processing (NLP) algorithms to identify relevant information in the inputted texts. Every SFL machine, on the basis of the respective pedagogical method, is focused on the extraction of some particular kind of textual information.

For the inquiry-based learning method, the SFL machine is focused on the extraction and classification of questions from texts, but also on the creation of new questions from declarative sentences. The SFL machine for the Dataset-based approach is focused on the identification of references to tables and images in the documents. Subsequently, said references are clustered in order to find homogeneous datasets to support learning methodologies. For a problem-based approach, the IT tool is specialized in obtaining a “map” of the inputted document set, by providing a list of the main topics of the inputted documents, organized in order to find specific sentences about a specific topic and the connection between similar documents. Finally, the SFL machine for scenario-based learning is focused on the extraction of pairs composed of questions and related answers.

Dive into the eSTEM project and SFL machines at