MACHINE LEARNING FOR DATA SCIENCE

The course builds on the knowledge acquired in previous courses on programming, statistics, data analysis, and introductory machine learning. It develops these foundations towards more advanced machine learning models and their use in data science applications.

The course provides students with the conceptual and operational tools needed to understand, design, train, evaluate, and use machine learning models. It gives particular attention to neural networks, model optimization, deep learning, generative models, and transformer-based models.

Theoretical concepts are introduced together with examples and applications, with the aim of developing the ability to reason about model behaviour, evaluation procedures, experimental choices, and the limitations of machine learning systems.

At the end of the course, students will have acquired knowledge and skills related to the main methods of advanced machine learning for data science applications, with particular attention to the design, training, evaluation, and critical use of models.

Students will be able to understand the general principles underlying advanced machine learning models, with particular reference to neural models, deep learning methods, and modern applications of artificial intelligence. They will also understand the main stages involved in the development of a machine learning system, from problem definition and choice of methodological approach to the evaluation of results.

Students will be able to apply machine learning methods to data science problems, use appropriate software tools, set up an experimental analysis, evaluate the results obtained, and discuss the limitations of the adopted approach.

Students will also be able to critically evaluate the quality of a model, the reliability of results, possible sources of error or bias, and aspects related to reproducibility, data quality, and the responsible use of artificial intelligence technologies.

Finally, students will be able to clearly present a machine learning project, justify the methodological choices adopted, discuss the results obtained, and independently study new methods, tools, and applications.

Students are expected to have basic knowledge of Python programming, linear algebra, probability, statistics, and data analysis.

Familiarity with the basic concepts of supervised and unsupervised machine learning, train/test splits, cross-validation, and model evaluation is recommended.

Basic experience with Python libraries such as NumPy, pandas, matplotlib, and scikit-learn is useful. No prior knowledge of deep learning frameworks is required.

The course introduces advanced machine learning methods for data science applications, with particular attention to neural models, deep learning methods, and modern applications of artificial intelligence.

After a review of the main methodological concepts needed to analyse, train, and evaluate models, the course focuses on the use of machine learning techniques for the development of predictive and generative systems.

The course will discuss aspects related to model design, optimization, performance evaluation, and interpretation of results. It may also introduce recent applications based on pre-trained models, generative models, and intelligent systems for the analysis of complex data.

Particular attention will be devoted to methodological and critical aspects related to the use of machine learning in applied contexts, including reproducibility, reliability, model limitations, data quality, bias, and responsible use of artificial intelligence systems.

Lecture notes, slides, notebooks, and selected readings will be made available through Moodle.

Documentation and tutorials from scikit-learn, PyTorch, Hugging Face, and related machine learning libraries will also be used. Additional references may be provided during the course for specific topics.

The assessment is based on a group project and an oral discussion.

The group project requires students to develop and document a project consistent with the topics of the course. The project aims to apply machine learning methods to a data science problem, including the definition of the problem, the choice of the methodological approach, data analysis, evaluation of the results, and discussion of the limitations of the work carried out.

The oral discussion is aimed at assessing the individual understanding of the topics covered in the course and of the choices made in the project. During the oral discussion, students may be asked to present the work carried out, justify the main methodological decisions, interpret the results obtained, and discuss possible improvements.

The final assessment will take into account the overall quality of the project, methodological correctness, clarity of presentation, critical analysis, and individual preparation demonstrated during the oral discussion.

oral

The lecturer has a duty to ensure that the rules regarding the authenticity and originality of exam tests and papers are respected. Therefore, if there is suspicion of irregular conduct, an additional assessment may be conducted, which could differ from the original exam description.

The assessment will take into account the level of understanding of the topics covered in the course, the ability to appropriately apply the methods studied, the quality of the project work, and the ability to critically discuss its results and limitations.

A sufficient grade will be awarded to students who demonstrate basic knowledge of the topics covered and adequate participation in the project work.

An intermediate grade will be awarded to students who demonstrate good understanding of the methods, ability to apply them correctly, and clarity in the presentation and discussion of the project.

A high grade will be awarded to students who demonstrate strong methodological understanding, autonomy in the analysis of the problem, quality in the development of the project, and critical ability in the discussion of results.

Honours will be awarded in the presence of excellent understanding of the topics, high-quality project work, methodological autonomy, and very clear communication skills.

The course combines lectures, examples, case studies, and computational activities. Depending on the topics covered, Python-based examples and exercises may be used to illustrate the implementation and evaluation of machine learning models.