- Assignment #1 consists in the prediction of default payments using a neural network. The dataset contains information on default payments, demographic factors, credit data, history of payment, and bill statements of credit card clients in Taiwan from April 2005 to September 2005.
- Assignment #2 consists on the prediction of grayscale images of letters P - Z. The goals are:
- The resolution of the problem of supervised classification with a traditional neural network (convolutional layers should not be used) with some motivations about the choices on data processing, number and dimension of the layers, optimization algorithms and regularization;
- The visual investigation of the reconstruction abilities of an auto-encoder architecture, i.e. comparison of the input and the reconstructed input;
- The use and evaluation of the encoded representation generated by the auto-encoder to solve the problem of supervised classification.
- The task of the Assignment #3 is the design of a CNN architecture and its training.
- The task of the Assignment #4 is Transfer Learning using a CNN pretrained on IMAGENET.
- The task of the Assignment #5 is Hyperparameter Optimization (HPO) of a neural network, with the aim to maximize its Accuracy on 10 fold cross validation using SMAC3.
⊜ Fabrizio D'Intinosante
- Cosa studio: Studente Magistrale di Data Science presso l'Università degli Studi di Milano-Bicocca;
- Studi precedenti: Laurea triennale in Economia e Statistica per le Organizzazioni presso l'Università degli Studi di Torino.