DSP422 Artificial Intelligence

Code DSP422
Name Artificial Intelligence
Status Compulsory/Courses of Limited Choice
Level and type Post-graduate Studies, Academic
Field of study Computer Science
Faculty
Academic staff Jānis Grundspeņķis
Credit points 4.0 (6.0 ECTS)
Parts 1
Annotation Artificial intelligence is developing towards four goals – to create systems that think or act like humans, as well as systems that think or act rationally. In this course students acquire knowledge about a modern approach to artificial intelligence – development of intelligent agents. The course is focused on properties, environment, architectures and programs of intelligent agents, logical agents, ontologies, planning, uncertain knowledge and reasoning, making simple and complex decisions, inductive learning, learning decision trees, neural networks and reinforcement learning. In development of a course work students must use their theoretical knowledge for implementation of agent based intelligent systems and analysis of their performance..
Goals and objectives
of the course in terms
of competences and skills
The goal of the course is to give theoretical knowledge and practical skills for development of agent-based intelligent computer systems.
Learning outcomes
and assessment
Students understand properties of intelligent agents, agent architectures and environments - Definitions of agents and their properties must be given and agent architectures as well as properties of environment must be explained in exam
Students understand agent structures and behaviour - The first laboratory task or the corresponding individual task
Students know structure of logical agents, knowledge representation and inference procedures - Perceptions and actions of logical agent must be defined and formalized, knowledge base must be constructed and inference procedure implemented in examination
Students know general-purpose ontologies and basic principles of their development - Basic principles of category representation must be described in examination
Students can apply various search algorithms - The second laboratory task or the corresponding individual task
Students understand structure of planning agent and representation of planning problem using formal language - The third laboratory task or the corresponding individual task
Students know methods of uncertain knowledge processing based on probabilistic reasoning - Bayesian network must be built; basic principles of utility theory must be defined and used for construction of decision network in examination
Students can apply modelling methods of simple and complex decision making agents - Utilities of states must be calculated for stochastic environment in examination
Students know essence of inductive learning and can apply methods of learning for decision trees and neural networks - The fourth laboratory task or the corresponding individual task
Students can apply algorithms of reinforcement learning - Set of learning examples must be generated and algorithms of reinforcement learning (passive reinforcement learning and temporal difference learning) must be used in examination
Students can apply value iteration and policy iteration algorithms - The fifth laboratory task or the corresponding individual task
Course prerequisites Basis strategies of state space search and knowledge representation schemas

[Extended course information PDF]