Infopaket

Artificial Intelligence Technologies

ОК 6

Обов’язкова дисципліна для ОП

Second

2021/2022

2 Semester

6

The student must know: the principles of engineering artificial intelligence based on intelligent agents, paradigms and basic tasks of machine learning, the mathematical model of an artificial neuron, the main functions of neuron activation, the principles of operation and learning of a multilayer perceptron and a convolutional neural network, methods of complexity regularization of neural networks, the operation principles of a support vector machine, performance metrics of predictive models, methodology for selecting the best model and hyperparameters. The student must have skills: to solve binary and multi-class classification problems using a multilayer perceptron and a support vector machine, to select the hyperparameters of classification models, to construct an image recognition model based on a convolutional neural network, and also to use pre-trained neural network models of modern architectures using transfer learning techniques.

Full-time form

The discipline “Artificial intelligence technologies” is based on such disciplines: “Programming”, “Higher Mathematics”, “Probability and Mathematical Statistics Theory”, “Algorithms and Methods of Calculation”, “Discrete Mathematics”, “Data Analysis with Python”.

The course "Artificial Intelligence Technologies" examines the basic approaches used to construct artificial intelligence systems, types of intelligent agents, machine learning paradigms and the main tasks solved by machine learning methods, a mathematical model of an artificial neuron, representation of neural networks with graphs, modern architectures of neural networks, logistic regression, single-layer and multilayer perceptrons, learning algorithms of artificial neural networks, generalization ability and performance metrics of predictive machine learning model, complexity regularization of neural networks, support vector machines, hyperparameter selection methodology, deep learning and convolutional neural networks.

1. S. Russell, P. Norvig. Artificial Intelligence: A Modern Approach. 4th Edition, 2020. – 1136 pages. 2. I. Goodfellow, Y. Bengio and A. Courville. Deep Learning. MIT Press, 2016. – 800 pages. 3. A. Gulli, S. Pal. Deep Learning with Keras. Packt Publishing, 2017. – 318 pages. 4. D. Sarkar, R. Bali, T. Sharma. Practical Machine Learning with Python: A Problem-Solver's Guide to Building Real-World Intelligent Systems. 1st Edition. Apress. 2017. – 555 pages. 5. A. C. Müller, S. Guido. Introduction to Machine Learning with Python: A Guide for Data Scientists. O'Reilly Media, 2016. – 398 pages.

Lectures, laboratory work, unsupervised work

Semester assessment: The academic semester has three semantic modules. The first two semantic modules are assessed after the completion of lecture topics 7 and 15 by compiling electronic tests. The third semantic module is assessed by the results of compilation (defense) of 4 laboratory works. Final assessment (in the form of exam): the exam form is written. The total result of exam is scored upto 40 points. The result of learning of the subject is considered positive if all 4 laboratory works is compilated (defensed) and the summary score for semantic modules and exam exceeds 60 points. Conditions of admission to the exam: the condition of admission to the exam is the student's receipt of a total of not less than the critical-calculated minimum of 20 points per semester.

Ukrainian