OVERVIEW

This course provides a comprehensive introduction to the fundamentals of Artificial Intelligence for biomedical and clinical applications. It is designed for students in Bioengineering and covers core machine learning and deep learning methods, along with practical guidance on building experimental pipelines that ensure robustness and reproducibility. Emphasis is placed on both theoretical foundations and hands-on practice, enabling students to apply data-driven approaches to real-world medical problems. The course integrates algorithmic principles with clinical relevance, fostering critical thinking about the opportunities, limitations, and ethical implications of AI in healthcare. Each class combines theoretical lectures with practical sessions to consolidate learning through direct experimentation.

AIMS AND CONTENT

LEARNING OUTCOMES

The course will introduce the fundamental concepts and principles of machine learning and artificial intelligence as it applies to medicine.

AIMS AND LEARNING OUTCOMES

Aims:

The course aims to:

• Introduce the main classes of AI techniques used in medicine, with a focus on machine learning.

• Provide hands-on experience in applying these techniques to clinical and biomedical datasets.

• Promote interdisciplinary literacy and a critical approach to integrating AI tools into healthcare systems.

At the end of the course, students will be able to:

1. Understand and explain the core concepts of machine learning and their relevance to medicine and biomedicine.

2. Design and implement supervised and unsupervised learning pipelines tailored to clinical datasets.

3. Apply feature selection and model evaluation techniques in the context of diagnostic and prognostic tasks.

4. Analyze the interpretability and trustworthiness of AI models in high-stakes decision-making.

5. Critically assess the ethical, legal, and social implications of AI in healthcare.

6. Collaborate in interdisciplinary teams and communicate results effectively to both technical and medical audiences.

PREREQUISITES

Recommended background knowledge

Probability, linear algebra, and statistics.

Calculus

Foundations in programming (preferably Python).

Prior exposure to biomedical signals or systems.

TEACHING METHODS

• Classes taught in person by the lecturer with the use of slides and/or blackboard.
• Hands-on practice using Jupyter notebooks.

SYLLABUS/CONTENT

Part 1 – Introduction to AI and Learning Paradigms

• Introduction to the course structure, objectives, and evaluation

• Fundamentals of Artificial Intelligence and its role in medicine

• Overview of data types and challenges in biomedical datasets

• Introduction to Python and Google Colab environment

• Supervised vs. Unsupervised learning: concepts and comparisons

• Introduction to bias, overfitting, and generalization

Part 2 – Machine Learning Methods for Prediction

• Core supervised learning techniques (e.g., decision trees, SVM, k-NN)

• Evaluation metrics: accuracy, ROC curves, confusion matrix

• Introduction to cross-validation and hyperparameter tuning

Part 3 – Biomarker Identification and Feature Selection

• Feature selection techniques: filter, wrapper, and embedded methods

• Regularization and dimensionality reduction

• Biomarker discovery from omics and high-dimensional data

Part 4 – Deep Learning Methods for Prediction

• Introduction to deep learning and neural networks

• Discussion on different architectures

Part 5 – Applications of AI in Medicine

• Case studies across radiology, genomics, and clinical decision support

• Ethical and regulatory challenges in clinical AI deployment

RECOMMENDED READING/BIBLIOGRAPHY

• Hastie, Trevor, et al. The elements of statistical learning: data mining, inference, and prediction. Vol. 2. New York: springer, 2009.
• Hastie, Trevor, Robert Tibshirani, and Martin Wainwright. "Statistical learning with sparsity." Monographs on statistics and applied probability 143 (2015): 143.
• Goodfellow, Ian, Yoshua Bengio, and Aaron Courville. Deep learning. MIT press, 2016.
• Topol, Eric. Deep medicine: how artificial intelligence can make healthcare human again. Hachette UK, 2019.

TEACHERS AND EXAM BOARD

LESSONS

LESSONS START

Classes usually start on the last week of September. The schedule of all courses can be found on EasyAcademy.

Class schedule

The timetable for this course is available here: Portale EasyAcademy

EXAMS

EXAM DESCRIPTION

The final exam consists of two components:

1. Quiz: A series of multiple-choice and short-answer questions assessing understanding of theoretical concepts. This quiz can be taken either in a single session or progressively throughout the course period, depending on student preference and class scheduling.

2. Written Test: A practical, open-ended exam where students are required to design and critically discuss machine learning pipelines applied to biomedical problems. Emphasis is placed on the ability to justify methodological choices, address issues of bias and generalization, and ensure interpretability and reproducibility.

ASSESSMENT METHODS

The evaluation process is composed of three components:

1. Test (Entry Quiz)
   A preliminary quiz on theoretical and practical course content. The test can be taken progressively during the teaching period or in a single session. Passing the test is mandatory to access the written exam.

2. Written Exam
   The main assessment consists of a written test where students are required to design and justify machine learning pipelines in response to applied biomedical problems. The written exam grade constitutes the base of the final grade.

3. Optional Project
   Students who achieve a grade of 28/30 or higher in the written exam and wish to improve their final grade may complete an individual project. The project involves solving a data analysis task (coding required) and can add or subtract up to 3 points from the written exam grade. The maximum final grade, including the project, is 30 e lode.