Unsupervised Domain Adaptation

This repository has been created for the Deep Learning project course delivered during my Master's program.

Introduction

In this repository we deliver a notebook to build, train and evaluate two different deep learning frameworks with respect to a baseline, that involves the topic of Unsupervised Domain Adaptation (UDA). For this assignment we use a UDA benchmark constisting of two domains, Product $P$ and Real World $RW$, treated as source domain and target domain, and viceversa. The aim of this project is to "propose a UDA technique to counteract the negative impact of the domain gap when training the model on the source distribution and evaluating it on the target distribution".

Note

To properly use the notebook, refere to the instructions available inside UDA.ipynb