Abstract:

When engineers translate research advances into society solutions, they need sensibility to all intercultural substrates of our society, avoiding leaving nobody in the gutter. Therefore, knowing different points-of-view when developing an artefact is of special interest. For that, the Collaborative Online International Learning (COIL) method offers a unique platform for their training.

In this work, we explain the experience of a COIL between Brazilian and Spanish computer science/engineering undergraduate students when dealing with healthcare applications. They were posed with the following research question: Are there any differences among the disease risk prediction of a model learnt from a data repository and people from different countries? What is the model dependency over the inter-subject variability?

To answer the question, 18 students from Spain (4 women) and 28 participants from Brazil (3 women) received a dataset to conduct an experimental study. Up to 12 teams of students, grouped according to individual affinities found after an ice-break session in which different questions about films and colors where answered, where set up, between 3 to 5 members each. Teams worked asynchronously, using different collaborative tools, providing a written report at the end. A final synthesis session was organized for students to be evaluated among each other.

Results told us insights both at the academic level and intercultural learning. For the former, where surprising outcomes were achieved, there were some drop outs that are maybe due by the idiom and elective nature of the course: the students focused on deploying a huge experimental machine learning work on the data provided, whilst the research question was poorly developed. The latter (31 valid responses) highlights the aspects of intercultural learning, that the students recognized to be learnt at the end of the journey, with a high satisfaction.

Keywords:

Engineering intercultural, Healthcare predictive models, Voice signal processing.