The teaching is organized as follows:

Learning objectives

The course is composed by two modules. The objective of the first, data management systems, is to allow students to acquire in-depth knowledge of the methodologies and tools necessary to manage large amounts of data in new systems not based on the relational model (we will therefore consider systems based on semi-structured or document-based models, NoSQL and extended models with time and space dimensions). In particular, the systems that must memorize data also produced by sensors and mobile devices will be considered, so that a correct integration of these new data sources with the corporate information system is possible. At the end of the course, the student will be able to design and query non-traditional databases with typical tools of the NoSQL approach. The objective of the second, "deep learning",is to provide the fundamentals of neural networks as evolution of linear models, including architecture, activation functions and backpropagation. He will know the basics of the optimization algorithms used in the training of neural networks. Furthermore, it will provide knowledge on basic neural network architectures: the student will know various types of neural networks such as convolutional neural networks (CNN) for image analysis, recurrent neural networks (RNN) for sequential data and transformers for natural language processing. He/she will be able to understand the design principles behind these architectures and their applications, and the underlying mathematical theories. He/she will also be able to apply the techniques to real-world problems, understanding what data inputs and outputs should be. The course will present also methods for data preprocessing and augmentation. In the second part the course will also present advanced neural network architectures, used to tackle natural language processing (NLP) and computer vision (CV) problems such as text classification, object recognition and machine translation. In this regard, the student will know how to use advanced tools such as variational encoders, generative adversarial networks, NERF, and large language models. Interpretability: The student will learn the basics of interpretability of a neural network. This will make it possible to offer guarantee tools on the results produced by neural networks. Deep Learning Frameworks and Tools: The student will have gained hands-on experience with popular deep learning frameworks such as TensorFlow, PyTorch, and Keras. You will be able to build, train, and evaluate deep learning models using these frameworks.