Predicting Generalization in Deep Learning via Metric Learning -- PGDL Shared task

The competition "Predicting Generalization in Deep Learning (PGDL)" aims to\nprovide a platform for rigorous study of generalization of deep learning models\nand offer insight into the progress of understanding and explaining these\nmodels. This report presents the solution that was submitted by the user\n\\emph{smeznar} which achieved the eight place in the competition. In the\nproposed approach, we create simple metrics and find their best combination\nwith automatic testing on the provided dataset, exploring how combinations of\nvarious properties of the input neural network architectures can be used for\nthe prediction of their generalization.\n