Deep networks for collaboration analytics: promoting automatic analysis of face-to-face interaction in the context of inquiry-based learning
Author
dc.contributor.author
Lamsa, Joni
Author
dc.contributor.author
Uribe, Pablo
Author
dc.contributor.author
Jiménez, Abelino
Author
dc.contributor.author
Caballero, Daniela
Author
dc.contributor.author
Hamalainen, Raija
Author
dc.contributor.author
Araya, Roberto
Admission date
dc.date.accessioned
2021-09-22T16:09:36Z
Available date
dc.date.available
2021-09-22T16:09:36Z
Publication date
dc.date.issued
2021
Cita de ítem
dc.identifier.citation
Journal of Learning Analytics Volume 8(1), 113–125. 2021
es_ES
Identifier
dc.identifier.other
10.18608/jla.2021.7118
Identifier
dc.identifier.uri
https://repositorio.uchile.cl/handle/2250/182056
Abstract
dc.description.abstract
Scholars have applied automatic content analysis to study computer-mediated communication in computer-supported collaborative learning (CSCL). Since CSCL also takes place in face-to-face interactions, we studied the automatic coding accuracy of manually transcribed face-to-face communication. We conducted our study in an authentic higher-education physics context where computer-supported collaborative inquiry-based learning (CSCIL) is a popular pedagogical approach. Since learners' needs for support in CSCIL vary in the different inquiry phases (orientation, conceptualization, investigation, conclusion, and discussion), we studied, first, how the coding accuracy of five computational models (based on word embeddings and deep neural networks with attention layers) differed in the various inquiry-based learning (IBL) phases when compared to human coding. Second, we investigated how the different features of the best performing computational model improved the coding accuracy. The study indicated that the accuracy of the best performing computational model (differentiated attention with pre-trained static embeddings) was slightly better than that of the human coder (58.9% vs. 54.3%). We also found that considering the previous and following utterances, as well as the relative position of the utterance, improved the model's accuracy. Our method illustrates how computational models can be trained for specific purposes (e.g., to code IBL phases) with small data sets by using pre-trained models.
es_ES
Patrocinador
dc.description.sponsorship
Academy of Finland
European Commission
292466
318095
Multidisciplinary Research on Learning and Teaching profiles I
ANID/PIA/Basal Funds for Excellence
FB0003