Objectives and outcomes
Learning about the ideas that have emerged during the research and development in artificial intelligence during its fifty-year development. Understanding how to build artificially intelligent entities. Upon completion of the course, students have a basic knowledge of artificial-intelligent systems (AI). They can determine what can be done with the AI approach. They can determine the problems that can be solved using AI. Students know the characteristics of AI methods. They can propose solutions to problems, and for some problems they can also choose and implement the appropriate AI method. They are prepared for narrower, professional subjects in the field of artificial intelligence and computer intelligence.
Lectures
The concept of artificial intelligence. An overview of AI fields and technologies. The real scope of artificial intelligence. Theoretical foundations of artificial intelligence. Cognitive psychology and neuroscience. Intelligent search. Mathematical, computational and statistical methods in artificial intelligence. Intelligent reasoning. Soft computing. Machine learning. Application of artificial intelligence. Intelligent data analysis. Intelligent agents. Multi-agent systems. Reflex agents. Agents with goals. Information extraction and information retrieval. Information extraction within natural language processing. Accuracy and responsiveness. Identification of named entities. A rule-based approach. An approach based on machine learning. Method of hidden Markov models. Maximum entropy method. Unsupervised information extraction. Finite-state transducers and their applications in the recognition of named entities and relations among them. Rule-based systems. The Semantic Web. Intelligent educational systems. Artificial intelligence technologies. Speech recognition and processing. Natural language processing. Robotics. Image processing. Neural networks.
Practical classes
Implementation of the search algorithms covered in lectures. Implementation of forward chaining and backward chaining reasoning rules. Software environments for the development of intelligent systems. Text segmentation and tokenisation tools. Working with open-source software spaCy. Examples of rule-based expert systems in different domains. Examples of problems solved by neural networks: classification, clustering, prediction, recognition, approximation and system modelling. Using open-source deep learning tools TensorFlow, Torch, Keras, Caffe. Areas of application of neural networks: medicine, finance, production, defence, social sciences.