Artificial intelligence 0800-SZIN

1. Introduction to AI methods: key issues of AI; definitions, status of AI; computational hardware, historical context, major discoveries, major projects, current situation.

2 Problem solving: search and optimisation algorithms. Representation of a problem in state space; reductive representation; search methods; heuristic search, examples of search-based programs; search vs. human thinking.

3. Knowledge representation and inference: types of knowledge; symbolic representation, predicate logic, fuzzy and approximate logic; rule systems; procedural representation.

4. Semantic networks and knowledge graphs; frames and scripts, program agents, search-based programs and their limitations, applications in games, from chess to StarCraft.

5. Decision support systems: knowledge acquisition; from robotic process automation to cognitive systems, design of advisory systems; application examples, large symbolic systems, CYC, SOAR and Watson.

6. Natural language understanding, symbolic approaches and sub-symbolic representations, non-ambiguity, semantics and the role of context, classification of utterances, machine translation, connectionism.

7 Data in AI systems: acquisition, cleaning, uncertainty, probabilistic distributions, Bayes' theorem, pattern recognition and correctness measures, vector representations - Word2Vec, Glove, attention mechanism.

8. Machine learning: types, supervised learning, decision trees, regression, SVM, similarity judgements, classification and regression using neural networks, associative memories, Hopfield model and Boltzmann machines.

9. Unsupervised learning: clustering, --average, hierarchical, data mining and visualisation, PCA, ICA, MDS, UMAP, t-SNE, manifolds.

10. Reinforcement learning, Markov processes, HMM, strategies and applications in games, robotics, reasoning.

11. Deep learning, convolutional network architecture, applications to image analysis.

12. Transformers, large linguistic and mulitmodal models, action models (Large Behavioral Models), generative artificial intelligence, GANs and other models.

13 Tools: HugginFace and other repositories, projects on GitHub, small local SLM models.

14. Human-machine interaction, models of mind, unified theories of cognition and autonomous LMMs, interpretability.

15. Global challenges, major AI systems, automation of creativity, discoveries, impact on society and economy.