The teaching is organized as follows:

Learning objectives

The course aims to provide the theoretical foundations and describe the main methodologies relating to the area of machine learning, together with the most recent techniques of deep learning. In particular, the course will deal with describing the methods of analysis, recognition and automatic classification of data of any type, typically called patterns. These disciplines are at the base, are used, and often complement many other disciplines and application areas of wide diffusion, such as computational vision, robotics, image processing, data mining, analysis and interpretation of medical and biological data, bioinformatics, biometrics, video surveillance, forecasting. More precisely, the methodologies that will be introduced in the course are often an integral part of the application areas mentioned above, and constitute the "intelligent" part of it with the final aim of understanding (classifying, recognizing, analyzing) the data coming from the process of interest (be they signals, images, strings, categorical, or other types). Starting from the type of data measured, the entire analysis pipeline will be considered, such as the extraction and selection of characteristics, supervised and unsupervised machine learning methods, parametric and non-parametric analysis techniques, and validation protocols, together with visualization necessary for understanding deep learning systems. At the laboratory level, real case studies and not just academic benchmarks will be presented, addressed with appropriate programming tools. In conclusion, the course aims to provide the student with a set of theoretical foundations and algorithmic tools to address the problems that may be encountered in strategic and innovative industrial sectors such as those involving the processing of large amounts of data (big data), multimedia, visual inspection of products, automation and forecasting.

Program

------------------------
UL: Foundation of Machine Learning - Theory
------------------------
The course is divided into two modules: Foundation of Machine Learning (ML) and Deep Learning (DL).
ML intends to provide the theoretical foundations and main methods relating to the analysis of data, not necessarily images. In short, theories and methods of statistical classification will be addressed. These topics are preparatory to the most recent Deep Learning techniques.
Addressed topics:
- Introduction: What is machine learning? Examples of Applications, main challenges of machine learning, tasks of machine learning, main ingredients
- Classification: binary classifier, performance measures (confusion matrix, precision, recall...), multi-class classification, multilabel classification, cross-validation
- Regression: linear regression, polynomial regression, logistic regression
- Bayesian Decision Theory and parameter estimation
- Nonparametric Methods: Histogram, Parzen windows, k-nearest neighbors
- Decision trees
- Ensemble learning and random forest
- Linear classifiers and discriminant functions: Perceptron, Relaxation, MSE, LMSE, gradient descent
- Linear transformations, Dimensionality reduction, Fisher transform. Principal Component Analysis, feature selection
- Kernel Methods and Support Vector Machines
- Unsupervised learning techniques: Clustering, gaussian mixture models
- Sequential data analysis: Markov models and hidden Markov models
- Machine learning versus deep learning
------------------------
UL: Foundation of Machine Learning - Laboratorio
------------------------
1) Introduction to Colab, Pytorch, Tensorflow
I/O data types, e.g., tabular data, images
2) Classification with scikitlearn: e.g., K-NN, evaluation
3) Data preparation, preprocessing, forward feature selection, data augmentation, normalization, missing data, one-hot vector
4) Principal component analysis & Fisher discriminant analysis
5) Clustering: K-means and elbow method, bag of words
6) Clustering methods and their comparisons, visualization methods (e.g., t-sne)
7) Support Vector Machines vs. Random forest
------------------------
UL: Deep Learning - Teoria
------------------------
The course is divided into two modules: Foundation of Machine Learning (ML) and Deep Learning (DL).
ML intends to provide the theoretical foundations and main methods relating to the analysis of data, not necessarily images. In short, theories and methods of statistical classification will be addressed. These topics are preparatory to the most recent Deep Learning techniques.
DL intends to provide theories and methods relating to the analysis of data (of various types, images, videos, text, sequences, etc.) using deep neural architectures, focusing on the structure and functioning of the different models such as, just as examples , convolutional networks, encoder-decoder models, attention models and transformer, and many others.
After an introduction of the importance of this area and its applications, the course includes topics such as artificial neural networks, convolutional networks, autoencoders - variational and non-variational, transformers, networks recurrent, generative models - adversarial and non-adversarial, multimodal models, methods for knowledge transfer and domain adaptation, etc.
The course will present the theoretical and methodological aspects, with associated application examples.
------------------------
UL: Deep Learning - Laboratorio
------------------------
The Lab classes are devoted to the development of algorithms in Python language of some of the models explained during the Theory classes.

Learning assessment procedures

------------------------
UL: Foundation of Machine Learning - Theory
------------------------
To pass the exam, students must demonstrate:
- Understanding of the principles underlying machine learning and methods for programming modules based on machine learning.
-Ability to articulate concepts of machine learning and programming of ML modules precisely and cohesively, without digressions.
-Application of acquired knowledge to solve practical problems presented through exercises, questions, and projects.
-The exam involves a project that can be performed individually or in pairs. The oral exam will cover questions related to the project, theoretical concepts, and lab exercises.
------------------------
UL: Foundation of Machine Learning - Laboratorio
------------------------
The exam includes a project that can be done individually or in pairs. The oral exam will cover project-related questions and laboratory exercises in accordance with the rules of the machine learning theory course.
------------------------
UL: Deep Learning - Teoria
------------------------
The exam will consist in the development of a project (2 people max, 3 people inexceptional cases, to be agreed with the teachers), followed by the writing of a technical report and an oral presentation.
During the oral presentation of the project, questions will be asked about the theoretical part of the course described in class, aimed at assessing the knowledge of the contents of the course.
------------------------
UL: Deep Learning - Laboratorio
------------------------
The exam will consist in the development of a project (2 people max, 3 people inexceptional cases, to be agreed with the teachers), followed by the writing of a technical report and an oral presentation.
During the oral presentation of the project, questions will be asked about the theoretical part of the course described in class, aimed at assessing the knowledge of the contents of the course.