Data-driven Health
Datadriven hälsa

BMEN35F, 7.5 credits

Valid from: Autumn 2022
Decided by: Professor Thomas Johansson
Date of establishment: 2022-03-24

General Information

Division: Biomedical Engineering
Course type: Course given jointly for second and third cycle
The course is also given at second-cycle level with course code: BMEN35
Teaching language: English

Aim

The course provides basic knowledge in the field of artificial intelligence and machine learning for applications in medicine and health. The course covers the chain from medical databases via algorithms to regulations and requirements for diagnostic software.

Goals

Knowledge and Understanding

For a passing grade the doctoral student must

• be able to and understand and have an overall picture of how machine learning and artificial intelligence can be designed for different medical issues.
• be able to apply the most common methods to real problems and real medical signals (Python)
• know and understand how special requirements and regulations for diagnostic systems and other medical devices affect the design and validation process.

Competences and Skills

For a passing grade the doctoral student must

• be able to understand the characteristics of the most common machine learning methods.
• be able to understand what settings need to be made for different methods.
• be able to understand how properties of medical training and validation data affect the performance of the methods.
• be able to choose the appropriate method for a described situation.
• be able to overview the consequences of different method choices and evaluation strategies that can be chosen for different types of problems
• be able to use available toolboxes for machine learning and in this way solve practical problems in Python.

Judgement and Approach

For a passing grade the doctoral student must

• have the ability to analyze and evaluate different machine learning algorithms, as well as interpret and describe their inherent principles.
• have insight into how the transparency and generalizability of different methods make them more or less suitable in different medical contexts.

Course Contents

Areas covered are: - Introduction of artificial intelligence in healthcare applications - Overview of machine learning algorithms and methods - How to choose ML methods for different applications - How to select settings and optimize performance - How to evaluate performance - Regulatory, social, ethical and legal issues regarding artificial intelligence in medicine -State-of-the-art AI that is applied to important medical fields such as ECG, neurology, biomedical imaging, heart sound, oncology, diabetes, etc. Practical work: - Introduction to Python / Jupyter / Colab (basics, linear algebra, plotting) - Linear models - Measurement values and visualization - Trees and knn - Ensemble methods - Neural networks (shallow, MLP, introduction to Keras / Tensorflow) - Deep Neural Networks (CNN) - Deep Learning (LSTM / RNN)

Course Literature

• Österberg, M. & Lindsköld, L.: AI for Better Health. 2020.
• Lindholm, A., Wahlström, N., Lindsten, F. & Schön, Thomas B.: Machine Learning - A First Course for Engineers and Scientists. Cambridge University Press, 2021.
• Bohr, A. & Memarzadeh, K.: Artificial Intelligence in Healthcare. Academic Press, 2020. ISBN 9780128184387.
• Xing, L., Giger, Maryellen L. & Min, J.: Artificial Intelligence in Medicine: Technical Basis and Clinical Applications. Academic Press, 2020. ISBN 9780128212592.

Instruction Details

Types of instruction: Lectures, seminars, exercises

Examination Details

Examination format: Written exam. Computer assignments
Grading scale: Failed, pass
Examiner:

Admission Details

Course Occasion Information

Start date: 2022-07-01
Course pace: Full time

Application Information

Apply by email to Course Coordinator

Contact and Other Information

Course coordinators:

• Christian Antfolk <christian.antfolk@bme.lth.se>
• Martin Stridh <martin.stridh@bme.lth.se>