Course detail
Artifical Inteligence
FSI-RAI Acad. year: 2027/2028 Summer semester
The course introduces fundamental approaches to artificial intelligence, including state-space search, stochastic optimization and machine learning, particularly the artificial neural networks including the convolutional and recurrent neural networks and transformers. The methods are demonstrated on simple engineering problems using appropriate tools (Matlab, TensorFlow, PyTorch).
Language of instruction
Czech
Number of ECTS credits
5
Supervisor
Entry knowledge
Vector and matrix operations, algorithm design skills, ability to implement algorithms in MATLAB and Python.
Rules for evaluation and completion of the course
The course grade is based on a functional software project that uses selected method of signal processing. Students must submit a report describing the task together with the source code and present the project to their classmates in a short presentation.
The report, in PDF format, must be at least 5 pages long and include the rationale for the assignment, detailed instructions, a description of the procedure, and the results achieved. The source code must be executable on the instructor’s computer. The presentation must be in PowerPoint or a compatible format.
Assessment is based on the fulfillment of the project’s objectives, as well as the quality of the code and text, taking into account the difficulty of the assignment.
Attendance at lectures is recommended, attendance at exercises is compulsory. The teacher specifies how students can substitute missed exercises.
Aims
Understanding of the basics of artificial intelligence approaches and the ability to apply them to engineering tasks.
Students will gain an overview of current artificial intelligence methods and will be capable to select and apply the appropriate methods to engineering problems.
The study programmes with the given course
Programme N-MET-P: Mechatronics, Master's, compulsory
Type of course unit
Lecture
26 hours, optionally
Syllabus
1. Introduction, areas of artificial intelligence, state space search, game theory – min/max algorithm
2. Evolutionary methods of state space search.
3. Basic paradigms of neural networks, unsupervised/supervised learning.
4. Backpropagation.
5. Approximation versus classification.
6. Convolutional neural networks – topology, convolution and pooling layers
7. Recurrent neural networks, LSTM
8. Reinforcement learning
9. Q-learning
10. Sarsa, DQN
11. Support Vector Machines
12. Transformers
13. Agent systems
Computer-assisted exercise
26 hours, compulsory
Syllabus
1. Essential tools: Matlab, Python, Tensor Flow, Keras, PyTorch.
2. Breadth/depth first search algorithms
3. Dijkstra algorithm, A-star
4. Min-max algorithm
5. Genetic algorithm
6. Layered networks, Neural Network Toolbox
7. Layered networks – examples
8. Convolution neural network – Tensor Flow, PyTorch
9. Reinforcement learning and Q-learning
10. Project, consultations
11. Project, consultations
12. Project, consultations
13. Project presentation