Course detail

Artificial Intelligence in Medicine

FEKT-AUINAcad. year: 2019/2020

The course focuses on the basic types of neural networks (with backpropagation, Hamming and Kohonen network). The second part focuses on the hierarchical and non-hierarchical cluster analysis. The third part is focused on the theory of fuzzy sets, fuzzy relations, fuzzy logic, fuzzy inference and approximate reasoning procedures. The following are the methods for relevant features selection and for evaluation of the results obtained by above tools of artificial intelligence.

Language of instruction

Czech

Number of ECTS credits

5

Mode of study

Not applicable.

Learning outcomes of the course unit

The graduate of the course is capable of:
- understanding principles of neural networks (with backpropagation errors, Hamming, Kohonen network),
- implementing the cluster analysis using the non-hierarchical or hierarchical methods,
- explaining the principle of fuzzy inference and approximate reasoning,
- performing selection of the most relevant features for further analysis,
- evaluating the performance of machine learning algorithms,
- giving examples of biomedical areas where above methods are widely used.

Prerequisites

The knowledge on the Bachelor´s degree level is requested, namely on numerical mathematics. The laboratory work is expected knowledge of Matlab programming environment.

Co-requisites

Not applicable.

Planned learning activities and teaching methods

Teaching methods include lectures and computer laboratories. Course is taking advantage of e-learning system. Students have to solve two tests (in the middle and at the end of semester, during the laboratories).

Assesment methods and criteria linked to learning outcomes

Requirements for completion of a course are elaborated every year by the lecturer responsible for the course.

- Total 25 points can be obtained for team project (non-zero point must be obtained)
- Total 75 points can be obtained for the exam (at least 12 points must be obtained from each of three parts of the exam)
- Total 50 points (for team project and the exam) must be obtained to pass the course

Course curriculum

1. Introduction to Artificial Intelligence.
2. Artificial neural networks, the neuron and its characteristics, the neuron as a classifier.
3. Learning the neuron with binary and real inputs and outputs, single-layer perceptron.
4. Multi-layer feed-forward network, the backpropagation algorithm.
5. Hamming network, Kohonen network.
6. Cluster analysis, hierarchical cluster analysis.
7. Non-hierarchical cluster analysis, k-means algorithm.
8. Fuzzy sets, fuzzy relations.
9. Logic, fuzzy logic, fuzzy inference, approximate reasoning.
10. Features selection and decorrelation.
11. Evaluation of classification, prediction and approximation algorithms.
12. Biomedical applications of machine learning tools.

Work placements

Not applicable.

Aims

Gaining knowledge about artificial neural networks, introduction to Hierarchical and non-hierarchical cluster analysis. Introduction to the theory of fuzzy sets, fuzzy relations, fuzzy logic, fuzzy inference methods as approximate reasoning. Furthermore, methods for relevant features selection and for evaluation of the results obtained using above approaches.

Specification of controlled education, way of implementation and compensation for absences

Delimitation of controlled teaching and its procedures are specified by a regulation issued by the lecturer responsible for the course and updated for every year (see Rozvrhové jednotky).
Basically:
- obligatory computer-lab tutorial (missed labs must be properly excused and can be replaced after agreement with the teacher)
- voluntary lecture.

Recommended optional programme components

Not applicable.

Prerequisites and corequisites

Not applicable.

Basic literature

Kozumplík, J., Provazník, I.: Umělá inteligence v medicíně. Elektronická skripta. ÚBMI FEKT VUT v Brně, Brno, 2007. (CS)

Recommended reading

Šnorek, M.: Neuronové sítě a neuropočítače. Skripta ČVUT, Praha, 2002 (CS)

Elearning

Classification of course in study plans

  • Programme BTBIO-A Bachelor's

    branch A-BTB , 3 year of study, winter semester, compulsory

  • Programme EEKR-CZV lifelong learning

    branch EE-FLE , 1 year of study, winter semester, compulsory

Type of course unit

 

Lecture

26 hod., optionally

Teacher / Lecturer

Syllabus

1. Introduction to Artificial Intelligence.
2. Artificial neural networks, the neuron and its characteristics, the neuron as a classifier.
3. Learning the neuron with binary and real inputs and outputs, single-layer perceptron.
4. Multi-layer feed-forward network, the backpropagation algorithm.
5. Hamming network, Kohonen network.
6. Cluster analysis, hierarchical cluster analysis.
7. Non-hierarchical cluster analysis, k-means algorithm.
8. Fuzzy sets, fuzzy relations.
9. Logic, fuzzy logic, fuzzy inference, approximate reasoning.
10. Feature methods for classification.
11. The logic of decision-making systems, predicate logic.
12. Inference and substantiation of claims.

Exercise in computer lab

26 hod., compulsory

Teacher / Lecturer

Syllabus

1. One artificial neuron without learning.
2. Using the neuron as a classifier.
3. Learning the neuron, δ-rule. Perceptron in Neural Network Toolbox.
4. The neural network without learning.
5. Neural network, the backpropagation (BP) algorithm.
6. BP feed-forward network in Neural Network Toolbox (example classification ECG cycles).
7. BP feed-forward network in Neural Network Toolbox (example approximation signals).
8. Hierarchical cluster analysis.
9. Non-hierarchical cluster analysis.
10. PCA decorrelation and symptoms.
11. Fuzzy cluster analysis.
12. Fuzzy inference and approximate reasoning.

Elearning