Caricamento in corso...

01QWYBH

A.A. 2024/25

Course Language

Inglese

Degree programme(s)

Master of science-level of the Bologna process in Ict For Smart Societies (Ict Per La Societa' Del Futuro) - Torino

Course structure

Teaching | Hours |
---|---|

Lezioni | 39 |

Esercitazioni in laboratorio | 21 |

Lecturers

Teacher | Status | SSD | h.Les | h.Ex | h.Lab | h.Tut | Years teaching |
---|---|---|---|---|---|---|---|

Visintin Monica | Professore Associato | IINF-03/A | 19,5 | 0 | 21 | 0 | 9 |

Co-lectures

Espandi

Riduci

Riduci

Teacher | Status | SSD | h.Les | h.Ex | h.Lab | h.Tut |
---|---|---|---|---|---|---|

Pagana Guido | Docente esterno e/o collaboratore | 19,5 | 0 | 0 | 0 |

Context

SSD | CFU | Activities | Area context |
---|---|---|---|

ING-INF/03 | 6 | B - Caratterizzanti | Ingegneria delle telecomunicazioni |

2024/25

The medium of instruction is English.
The objectives of this course are to use machine learning in public health applications, in particular in the areas of basic research, prevention, diagnostic process, management of elderly people at home. The course is designed jointly with the course "Statistical Learning and Neural Networks", with the objective to provide students with a coordinated "machine learning" approach that can be applied to several ICT problems; in particular, "Statistical Learning and Neural Networks" deals primarily with machine learning in terms of classification and neural networks, while "ICT for health" addresses regression and clustering topics. Some classification techniques not analyzed in "Statistical Learning and Neural Networks" are analyzed in "ICT for health".
The course is divided into two parts: 1) the description of some of the many health issues and 2) the description and use of the machine learning techniques that can be used to solve these issues.
Several laboratory experiences are included, and the knowledge of the health issues from the medical point of view is fundamental for the correct system implementation. Python will be used as programming language (in particular Pandas and Scikit-learn ) and a learn-by-doing approach will be used.

The objectives of this course are to utilize machine learning in public health applications, particularly in the areas of basic research, prevention, diagnostic processes, and the management of elderly individuals at home. The course is designed in conjunction with the "Statistical Learning and Neural Networks" course, with the aim of providing students with a cohesive "machine learning" approach that can be applied to various ICT (Information and Communication Technology) problems. Specifically, "Statistical Learning and Neural Networks" primarily focuses on machine learning in terms of classification and neural networks, while "ICT for health" addresses regression and clustering topics. Certain classification techniques that are not covered in "Statistical Learning and Neural Networks" are discussed in "ICT for health."
The course is divided into two parts: 1) the explanation of several health issues and 2) the description and utilization of machine learning techniques to address these issues. Various laboratory experiences are included, where a solid understanding of the health issues from a medical perspective is essential for proper system implementation. Python, particularly Pandas and Scikit-learn, will be used as the programming language, and a learn-by-doing approach will be employed.

Knowledge of:
- basics in some health issues (management of elderlies, Parkinson's disease, EEG, ECG, dermatology, etc)
- e-health and m-health applications
- telemedicine applications
- regression techniques
- clustering techniques
- classification techniques
Ability to:
- understand the issues of a telemedicine application
- apply regression techniques in health problems
- apply clustering techniques in health problems
- apply classification techniques in health problems
- use open-source machine learning software

In this course, you will develop the following abilities:
Understand the issues related to an e-health application.
Explain the differences among AI, ML, Deep Learning.
Explain the regression technique covered in the lectures and apply them to health problems.
Explain the clustering techniques discussed in the lectures and apply them to health problems.
Explain the classification techniques presented in the lectures and apply them to health problems.
Compare the various clustering techniques described and choose the most appropriate one for a given problem.
Compare the different classification techniques described and determine the best one to address a given problem.
By the end of the course, you will have the knowledge and skills necessary to understand e-health applications, apply regression, clustering, and classification techniques to health-related problems, compare and select the most suitable techniques for specific problems.

Knowledge of probability theory, linear algebra, optimization techniques

Basic knowledge of probability theory, linear algebra, optimization techniques, programming.

- Description of some e-health, m-health, and telemedicine applications (2 CFU) on the following topics: smart aging, fitness, Parkinson's disease, EEG, ECG, dermatology, lean in health care (2.1 CFU).
- Review of linear algebra and basics on optimization methods (0.6 CFU).
- Introduction to Python (0.3 CFU).
- regression techniques: linear regression, PCR, Gaussian processes for regression (0.9 CFU)
- clustering techniques: k-means, hierarchical trees, DBSCAN (0.9 CFU)
- classification techniques: decision trees and information theory, Hidden Markov Models (0.9 CFU)
- Independent component analysis (ICA) applied to EEG (0.3 CFU)

- Description of some e-health, m-health, and telemedicine applications (2.1 CFU) on the following topics:
- smart aging
- fitness
- Parkinson's disease
- EEG
- ECG
- dermatology/teledermatology
- lean in health care
- management of emergencies
- differences among AI (Artificial Intelligence), ML (Machine Learning) and Deep Learning..
- Review of linear algebra and basics on optimization methods (0.6 CFU).
- Introduction to Python (0.3 CFU).
- Regression techniques: linear regression and tests of Gaussianity, applied to score of Parkinson's disease (0.9 CFU)
- Clustering techniques: k-means, hierarchical trees, and DBSCAN, applied to body images like CTScan or pictures of skin moles (0.9 CFU)
- Classification techniques: sensitivity, specificity, prevalence, incidence applied to tests based on blood markers; decision trees and information theory applied to detection of an illness like Chronic Kidney Disease (0.9 CFU)
- Independent component analysis (ICA) applied to EEG (0.3 CFU).

Lectures will describe the health context and the problem to be solved, then the relevant ICT/learning machine methods to be used to solve the problem are discussed and implemented in Python in the laboratory classes.

During the lectures, the course will provide descriptions of the health context and the specific problems that need to be addressed (20 hours). Subsequently, the relevant ICT methods to solve these problems will be discussed (20 hours), and practical implementation using Python will be carried out in the laboratory classes (20 hours). While no formal report on the lab activities is required, it is crucial to actively participate in the lab exercises as they are fundamental for understanding and correctly applying the methods explained in the lectures. Students have the freedom to work in groups during the lab sessions, encouraging collaborative learning and problem-solving. Questions on the lab activities might be asked at the exam.

- Class slides will be available on the portal
- K. Murphy, "Machine Learning, a probabilistic perspective", MIT press, 2012
- Christopher M. Bishop, "Pattern Recognition and Machine Learning", Springer-Verlag New York, 2006
- David J.C. MacKay, "Information Theory, Inference and Learning Algorithms"
Cambridge University Press 2003
- C.E. Rasmussen, C.K.I. Williams, Gaussian Processes for Machine Learning, the MIT Press, 2006

Reference material: class slides for both lectures and laboratories.
Additional material:
- K. Murphy, "Machine Learning, a probabilistic perspective", MIT press, 2012
- Christopher M. Bishop, "Pattern Recognition and Machine Learning", Springer-Verlag New York, 2006
- David J.C. MacKay, "Information Theory, Inference and Learning Algorithms"
Cambridge University Press 2003
- C.E. Rasmussen, C.K.I. Williams, Gaussian Processes for Machine Learning, the MIT Press, 2006

Slides; Esercitazioni di laboratorio; Video lezioni tratte da anni precedenti;

Lecture slides; Lab exercises; Video lectures (previous years);

...
The student must write two reports on the lab activity; together with the report the student must provide the zipped folder with the Python scripts. Each of these reports gets a grade between 0 and 5 (0 if the report is missing), for a total maximum grade equal to 10; the grade depends on the correctness of the results and on the completeness and clearness of the document.
The mandatory oral exam consists of 3 questions, 1 question about the health issues, i.e. Parkinson's disease, dermatology, etc. as described in the lectures, 2 questions about the algorithms and methods described in the lectures and the lab activity. The student gets a grade from 0 to 7 on each question; the grade depends on the ability to describe and critically discuss the learned methods.
The grades of the reports and the oral exam are added together to obtain the final grade. The "lode" is given to students with an overall grade 31.
The ability of the student to apply the described machine learning techniques in Python will be checked through the analysis of the report and the Python scripts. The knowledge of the health issues with possible solutions and the knowledge of the regression, clustering and classification techniques will be checked during the oral exam. The student will improve his/her soft-skills related to the ability of writing a technical report, and the ability to discuss ideas during the oral exam.

Gli studenti e le studentesse con disabilità o con Disturbi Specifici di Apprendimento (DSA), oltre alla segnalazione tramite procedura informatizzata, sono invitati a comunicare anche direttamente al/la docente titolare dell'insegnamento, con un preavviso non inferiore ad una settimana dall'avvio della sessione d'esame, gli strumenti compensativi concordati con l'Unità Special Needs, al fine di permettere al/la docente la declinazione più idonea in riferimento alla specifica tipologia di esame.

The mandatory oral exam for this course comprises three questions. The first question focuses on health issues covered in the lectures, such as Parkinson's disease, dermatology, and others. The second and third questions pertain to the algorithms and methods discussed in the lectures and applied in the lab activities. The questions will check that the student has acquired the abilities described in the expected learning outcome section (ability to explain, to compare, to understand). The oral exam evaluates the student's understanding of health issues and potential solutions, as well as their grasp of regression, clustering, and classification techniques. Furthermore, the exam aims to improve the student's soft skills, particularly their ability to engage in discussions, articulate ideas, and describe methods effectively.
Each question is graded on a scale of 0 to 10, considering factors such as the accuracy of the answer, response time, and the student's ability to critically discuss the learned methods.
The "lode" distinction is awarded to students who achieve a grade of 30 and demonstrate exceptional confidence in their knowledge, promptness in answering questions, clarity in their responses, and the ability to apply learned methods to new scenarios.
Access to the oral exam is possible only for those students who pass a closed book short quiz (15 minutes) with 9 multiple choice questions (at least 5 out of 9 points, wrong answers give a negative contribution), administered through the PoliTO Exam platform, the same day as the oral exam, same room. The quiz grade is not used for the final grade.

In addition to the message sent by the online system, students with disabilities or Specific Learning Disorders (SLD) are invited to directly inform the professor in charge of the course about the special arrangements for the exam that have been agreed with the Special Needs Unit. The professor has to be informed at least one week before the beginning of the examination session in order to provide students with the most suitable arrangements for each specific type of exam.