Saúl Calderón participates in the paper Dealing with Distribution Mismatch in Semi-supervised Deep Learning for Covid-19 Detection Using Chest X-ray Images: A Novel Approach Using Feature Densities
In this work, is assessed the impact of the distribution mismatch between the labelled and the unlabelled datasets, for a semi supervised model trained with chest X-ray images, for COVID-19 detection. Under strong distribution mismatch conditions, was found an accuracy hit of almost 30%, suggesting that the unlabelled dataset distribution has a strong influence in the behavior of the model.
In the paper is proposed a straightforward approach to diminish the impact of such distribution mismatch. The method uses a density approximation of the feature space. It is built upon the target dataset to filter out the observations in the source unlabelled dataset that might harm the accuracy of the semi-supervised model.
The article has been accepted in the Applied Soft computing journal (Q1, Impact Factor 7.7). Professor Saúl Calderón, member of the PARMA group, participates in this paper. To read the complete preprint you can visit the following link: https://arxiv.org/abs/2109.00889