|
|
Dettaglio Insegnamento
| Anno Accademico |
2022/2023 |
| Corso di Studi |
Dott. - MI (1373) Bioingegneria / Bioengineering |
| Anno di Corso |
1 |
| Codice Identificativo |
059008 |
| Denominazione Insegnamento |
AI METHODS FOR BIOENGINEERING CHALLENGES |
| Tipo Insegnamento |
MONODISCIPLINARE |
| Crediti Formativi Universitari (CFU) |
5.0 |
| Programma sintetico |
MISSION AND GOALS
The course mission is to illustrate to the students timely AI techniques and have them practice on clinical data challenges in the biomedical signal and image processing domain. The challenge format, very popular in the last years within international meetings, makes available large free-of-use datasets usable to develop and test innovative solutions based on AI tools. The goals are: 1) make students familiar with modern AI solution leveraging on artificial-neural-networks; 2) present the python-based programming environment with distributed computing capability; 3) enable students to apply general AI tools to solve a specific problem in biomedical challenges
SUBJECT AND PROGRAMME OF THE COURSE
General overview of biomedical challenges in the areas of: 1) oncological diagnosis and pathological staging; 2) organ and structure segmentation on images; 3) bio-signal analysis and event detection; 4) 3D protein structure and folding prediction. Description of the challenges and data repositories like Physionet/PhysioBank, National Cancer Archive, MICCAI challenges, Google Deepmind Alphafold2. For each of the described challenges, the most advanced AI solutions for processing and analyzing biomedical data,
reported in the recent literature, will be detailed. Such solutions will include, but not limited to, convolutional encoding-classification networks for oncological diagnosis, 2D/3D U-net architectures for organ/structure segmentation, recurrent and transformer network for bio-signal analysis, graph-networks for protein folding prediction. Experimental sessions will be started by splitting the students in teams who will then develop their own AI solution for the selected challenge.
TEACHING ORGANIZATION
The organization of the course activities will adopt multiple teaching strategies including 1) frontal class for presentation of the challenges and related techniques; 2) hands-on programming sessions (methods and processing environment); 3) team-working on thematic challenge applications. The course lecturers will be available for tutoring/mentoring
during team-working activities.
TEACHING MATERIAL
Lecture slides and video recordings Online videos SW libraries Example notebooks
LEARNING EVALUATION
Students, split in teams, will be asked to deliver a proceedings-like report on the assigned challenge. The final score will be attributed on the basis on the originality, methodological correctness and performance of the solution. Authors of the best solution will be encouraged and supported to submit a contribution to a sector conference. |
| Settori Scientifico Disciplinari (SSD) |
--
|
|
Scaglione
|
Docente
|
Programma dettagliato
|
|
Da (compreso)
|
A (escluso)
|
|
A
|
ZZZZ
|
Cerveri Pietro, Barbieri Riccardo, Mainardi Luca
|
|
|
