Minor specialization in Artificial Intelligence (AI) is a suitable complement to all fields of the Master's program. Minor in AI counts if the student finishes any four compulsory courses in the field of AI. At least 3 of them must not be among the compulsory courses of the student's main specialization. AI courses are divided primarily into those focused on 1) machine learning and data analysis: Statistical Machine Learning, Symbolic Machine Learning or 2) algorithmic artificial intelligence: Multi-agent Systems, Planning for Artificial Intelligence, Logical Reasoning and Programming, and Artificial Intelligence in Robotics. While for sourses such as data science, computer vision, digital painting or bioinformatics, algorithmic artificial intelligence is a suitable complement; machine learning courses are more suitable for cyber security. Other fields can be complemented by any combination of the courses.
- Prof. Dr. Michal Pěchouček, MSc.
Passage through study
Minor in AI counts if the student finishes any four compulsory courses in the field of AI. At least 3 of them must not be among the compulsory courses of the student's main specialization.
General characteristics of the courses:
- Course Statistical Machine Learning (BE4M33SSU) aims to develop systems (models and algorithms) able to learn to solve tasks given a set of examples and some prior knowledge about the task. The course will cover the following topics - Empirical risk minimization, consistency, bounds - Kernel SVMs, RKHS, regression - Semi-supervised learning - Unsupervised learning, EM algorithm, mixture models - Bayesian learning - Deep (convolutional) networks and Boltzmann machines (graphical models) - Supervised learning for deep networks - Hopfield nets and energy minimisation (MAP in MRFs) - Structured output SVMs - Sampling methods, sampling from models - Ensemble learning, random forests.
- Course Computational Game Theory (B4M36MAS) provides an introduction to concepts, models, and algorithms for autonomous agents and multi-agent systems. The first part of the course introduces single-agent models and control architectures; the second part explains key multi-agent models and algorithms, both for cooperative and non-cooperative multi-agent settings. Upon successful completion of the course, students will be able to understand main multi-agent concepts, be able to map real-world multi-agent problems to multi-agent formal models and apply algorithmic techniques to solve them. .
- Course Planning for Artificial Intelligence (B4M36PUI) covers the issues of planning in artificial intelligence. It focuses mainly on domain-independent models of planning problems: state-space planning, plan-space planning, heuristic planning, graph planning or hierarchical planning. Students will also be acquainted with the issue of planning under uncertainty and planning treated as a decision-making task (MDP and POMDP).
- Course Symbolic Machine Learning (B4M36SMU) explains methods through which an intelligent agent can learn, that is, improve its behavior by interacting with the environment. The learning scenarios will include concept learning: we will study online learning and batch learning from i.i.d. data. We will define the mistake-bound and PAC model of learning. Strong emphasis will be on logical representations of learned knowledge, including operators for generalization of logic clauses. Learning probability distributions with a graphical model (Bayes Networks) Reinforcement learning Universal learning with the Kolmogorov prior. Time permitting, we will also discuss active learning with queries.
- Course Artificial Intelligence in Robotics (B4M36UIR) aims to acquaint students with the use of planning approaches and decision-making techniques of artificial intelligence for solving problems arising in autonomous robotic systems. Students in the course are employing knowledge of planning algorithms, game theory, and solving optimization problems in selected application scenarios of mobile robotics. Students first learn architectures of autonomous systems based on reactive and behavioral models of autonomous systems. The considered application scenarios and robotic problems include path planning, persistent environmental monitoring, robotic exploration of unknown environments, online real-time decision-making, deconfliction in autonomous systems, and solutions of antagonistic conflicts. In laboratory exercises, students practice their problem formulations of robotic challenges and practical solutions in a realistic robotic simulator or consumer mobile robots.
- Course Logical Reasoning and Programming (B4M36LUP) aims to explain selected significant methods of computational logic. These include algorithms for propositional satisfiability checking, logical programming in Prolog, and first-order theorem proving and model-finding. Time permitting, we will also discuss some complexity and decidability issues pertaining to the said methods.