Now showing 1 - 10 of 14
No Thumbnail Available
Publication

Real Estate Monitoring System Based on Remote Sensing and Image Recognition Technologies

2017, Kodors, Sergejs, Aldis Rausis, Aivars Ratkevics, Janis Zvirgzds, Teilāns, Artis, Ivonna Ansone

Geoinformation are changing fast, therefore a change detection of real estate must be processed in short time. The increasing resolution of sensed geospatial data creates critically important to develop high performance computing solutions to process geospatial information. The topic of scientific work is the real estate monitoring system based on image recognition and remote sensing technologies. System's practical application is automatic building recognition from LiDAR data using saliency based method, vector map generation and change detection in actual cadastral maps. The scientific work describes high performance computing solution and gives its performance comparison with traditional method.

No Thumbnail Available
Publication

AUTOMATION OF FEEDBACK ANALYSIS IN ASYNCHRONOUS E-LEARNING

2023, Rohit, Grabusts, Pēteris, Teilāns, Artis, Atis Kapenieks

With the recent hit of the Pandemic study process is shifted to E-learning. Measuring the actual progress of a student in Asynchronous e-learning because of machine-human interaction and feedback is also considered a primary issue in this area. With the rapid development of artificial intelligence, computers can capture surroundings. Image processing is a rising technique in Artificial Intelligence (AI).  Recognition of an individual's emotion helps identify the person's inner state. It is easy to measure a student's feedback about the session by doing it. The main functionality of this project is to capture the student's enough frames during the study session and provide the analysis of the average emotions to the administration panel. This prototype was used on 10 minutes of lecture to capture the emotion. The primary goal of this proposed system is to capture the learner's emotion at a specific interval during the e-learning session and provide the feedback of it to the instructor.

No Thumbnail Available
Publication

Mathematical Model of the Distribution of Laser Pulse Energy

2016, Pavels Narica, Teilāns, Artis, Lazov, Lyubomir, Pavels Cacivkins, Teirumnieks, Edmunds

Method allows for modelling of the complex process of laser pulse energy distribution over flat work surface. The process of calculating the correct result does not use common lasing formulas but instead employs the mathematical model of matrix multiplication of three input matrices representing a pulse model, a line model, and a plane model. The pulse model represents the distribution of planar energy densities within the laser pulse. The line model represents the distribution of pulses within the line. The plane model represents the distribution of lines within the plane. Because mathematical model is implemented within a spreadsheet processor, its size can be adjusted as needed and it can be instantiated multiple times for simultaneous modelling of different input parameters.

No Thumbnail Available
Publication

Methodology for Similarity Assessment of Relational Data Models and Semantic Ontologies

2017-01-17, Zarembo, Imants, Teilāns, Artis, Barghorn, Knut, Merkuryev, Yuri, Berina, Gundega

In the upcoming age of semantic web there is a large number of relational databases being widely used. When time comes for a legacy relational database to migrate to semantic web or to be integrated with it, an important issue of determining similarity (compatibility) between two data models expressed in different ways arises. The goal of this paper is to describe the methodology for similarity assessment of relational database models and semantic data models and to present an ontology matching tool research prototype. The methodology consists of a set of steps, including transformation rules for data models, whose compatibility must be assessed, to the same ontology representation and applying ontology matching techniques. The methodology enables domain experts to perform a matching task semi-automatically between a relational data model and data model expressed as an ontology. The results of the semi-automatic matching are manually verified by the domain experts. The methodology was approbated using a use case from land administration domain. In the use case compatibility of data model provided by an international standard and a relational database had to be assessed.

No Thumbnail Available
Publication

INFORMATIONAL WARFARE – INFLUENCE ON INFORMATIONAL STRUCTURES

2019, Grabusts, Pēteris, Zorins, Aleksejs, Teilāns, Artis

The concept of information warfare encompasses the use of information and communication technologies to gain an advantage over a potential opponent. The information warfare is the manipulation with the information that trusts the goal, so that the goal should make decisions about its interests in the interests of opponents. Information structures are treated as systems that process different types of information, provide storage and access to users. Such structures may enclose neural networks, self-learning systems etc. They need to be ready to learn, respond to threats and ensure their safety, which is topical in today's information warfare. This paper will address aspects related to the security of information systems from a system theory point of view. The knowledge base of information structures can be elements of artificial intelligence, which security must be protected against various threats. The authors considers artificial neural networks to be one of the potential threats in the context of information warfare.

No Thumbnail Available
Publication

METHOD FOR COLOR LASER MARKING PROCESS OPTIMIZATION WITH THE USE OF GENETIC ALGORITHMS

2017, Pavels Narica, Lazov, Lyubomir, Teilāns, Artis, Grabusts, Pēteris, Teirumnieks, Edmunds, Pavels Cacivkins

Optimization of color laser marking process mostly depends on effective identification of optimal values of laser marking parameters. This is a difficult combinatorial optimization problem, which is still essential for companies that use laser marking systems. The study proposes a new approach to the process optimization through the use of genetic algorithms, carrying out preliminary experimental investigation, analyzing the laser marking results, and presenting possible improvements to the current implementation of genetic algorithms.

No Thumbnail Available
Publication

Assessment of Name Based Algorithms for Land Administration Ontology Matching

2015, Zarembo, Imants, Teilāns, Artis, Aldis Rausis, Jazeps Buls

Abstract The purpose of this paper is to tackle semantic heterogeneity problem between land administration domain ontologies using name based ontology matching approach. The majority of ontology matching solutions use one or more string similarity measures to determine how similar two concepts are. Due to wide variety of available general purpose techniques it is not always clear which ones to use for a specific domain. The goal of this research is to evaluate several most applicable string similarity measures for use in land administration domain ontology matching. To support the research ontology matching tool prototype is developed, where the proposed algorithms are implemented. The practical results of ontology matching for State Land Service of Latvia are presented and analyzed. Matching of Land Administration Domain Model international standard, present Latvian Land Administration ontology is conducted.

Thumbnail Image
Publication

E-LEARNING: DEVELOPING TOMORROW'S EDUCATION

2021, Rohit, Grabusts, Pēteris, Teilāns, Artis

E-learning refers to the term to deliver education or training using digital resources. Computer-based learning, which is considered the keystone of today's E-learning concept, was born in the 80s. Earlier E-learning provides education using only text as with the development in technology it allows adding various forms, i.e., Graphical Text, Images, Video Conferencing etc. In today's time, the concept of is E-learning growing at a rapid pace. Improved bandwidth and growing technology helped in pushing the expansion of E-learning. Along with the university, large corporate companies are also resorting to E-learning. E-learning provides many advantages as compare to Instructor-led training (ILT). E-learning saves the times of travel as physical presence is not required. Education can be provided from anywhere at any time. E-learning is cost-effective also as the course, once developed, can be modified easily. There can some concern which can be faced by the trainer and leaners in future. Adopting E-learning will be a step towards saving the environment. It will be environmentally friendly as tablets will replace books; paper notes will be replaced with digital messages. Digital tools will help to reduce the burden of a student. Artificial Intelligence is a prevalent concept in computer science. A branch of AI, known as a Neural Network, is based on the human brain. The research's main aim is to review existing methods and analyse further possibilities of E-learning systems with neural networks.

No Thumbnail Available
Publication

Machine Learning Technology Overview In Terms Of Digital Marketing And Personalization

2021, Teilāns, Artis, Nikolajeva, Anna

The research is dedicated to artificial intelligence technology usage in digital marketing personalization. The doctoral theses will aim to create a machine learning algorithm that will increase sales by personalized marketing in electronic commerce website. Machine learning algorithms can be used to find the unobservable probability density function in density estimation problems. Learning algorithms learn on their own based on previous experience and generate their sequences of learning experiences, to acquire new skills through self-guided exploration and social interaction with humans. An entirely personalized advertising experience can be a reality in the nearby future using learning algorithms with training data and new behaviour patterns appearance using unsupervised learning algorithms. Artificial intelligence technology will create website specific adverts in all sales funnels individually.

No Thumbnail Available
Publication

A Metamodel Based Approach for UML Notated Domain Specific Modelling Language

2011, Arnis Kleins, Teilāns, Artis, Yuri Merkuryev, Ojars Krasts

This paper focuses on a metamodel based approach to Unified Modelling Language (UML systems modelling and simulation. The approach allows creating a system model by operating with artefacts from the problem domain. As a novelty for UML modelling, especially for simulation purposes, the presented meta-model is enriched by a set of stochastic attributes of modelled activities. Modelling process is ensured by developing UML based Domain Specific Language (DSL) that is suitable for the metamodel, where UML diagrams are complemented with attributes necessary for model simulation. A modelling tool prototype was developed with Microsoft Visual Studio using Microsoft Visualization and Modelling SDK. Elaborated models are stored in a modelbase which conforms to the described metamodel. Relevant DEVS simulation software will be developed for ability to run those models and analyse gathered results. The given approach facilitates increases of the productivity in development of domain specific modelling and simulation tools up to 10 times.