Facebook Instagram Twitter RSS Feed Back to top

Information Page of SAS Organisation


Institute of Informatics

International Projects

COBRA - COnversational BRAins


Duration: 1. 2. 2020 - 31. 1. 2024
Program: Horizont 2020
Project leader: Prof.Mgr. Beňuš Štefan PhD.
Annotation:The EU-funded COBRA project will train the next generation of researchers to accurately characterise and model the linguistic and cognitive brain mechanisms that allow conversation to unfold in both human-human and human-machine interactions. The first challenge will be to determine how alignment and prediction may both rely on and contribute to setting up brain-to-brain coupling relationships. The second will relate to the development of computational models of alignment and prediction for more socially-acceptable text-to-speech synthesisers, human-machine dialogue systems, and social robots.
Project web page:https://www.cobra-network.eu/the-project/

EGI-ACE - EGI Advanced Computing for EOSC (EGI-ACE)

EGI pokročilé počítanie pre EOSC

Duration: 1. 1. 2021 - 30. 6. 2023
Program: Horizont 2020
Project leader: doc. Ing. Hluchý Ladislav CSc.
Annotation:The mission of the EGI-ACE project of EGI is to implement the Compute Platform of the European Open Science Cloud, by delivering a secure federation of Cloud compute and storage facilities in collaboration with providers of the EGI Federation, commercial providers, data providers and international research infrastructures of pan-European relevance.
Project web page:https://www.egi.eu/projects/egi-ace/

EOSC-Synergy - European Open Science Cloud - Expanding Capacities by building Capabilities (EOSC-Synergy)

Európsky cloud pre otvorenú vedu – rozšírenie kapacít budovaním infraštruktúrneho potenciálu

Duration: 1. 9. 2019 - 31. 10. 2022
Program: Horizont 2020
Project leader: doc. Ing. Hluchý Ladislav CSc.
Annotation:EOSC-synergy extends the EOSC coordination to nine participating countries by harmonizing policies and federating relevant national research e-Infrastructures, scientific data and thematic services, bridging the gap between national initiatives and EOSC. The project introduces new capabilities by opening national thematic services to European access, thus expanding the EOSC offer in the Environment, Climate Change, Earth Observation and Life Sciences. This will be supported by an expansion of the capacity through the federation of compute, storage and data resources aligned with the EOSC and FAIR policies and practices. EOSC-synergy builds on the expertise of leading research organizations, infrastructure providers, NRENs and user communities from Spain, Portugal, Germany, Poland, Czech Republic, Slovakia, Netherlands, United Kingdom and France, all already committed to the EOSC vision and already involved in related activities at national and international level. Furthermore, we will expand EOSC’s global reach by integrating infrastructure and data providers beyond Europe, fostering international collaboration and open new resources to European researchers. The project will push the EOSC state-of-the-art in software and services life-cycle through a quality-driven approach to services integration that will promote the convergence and alignment towards EOSC standards and best practices. This will be complemented by the expansion of the EOSC training and education capabilities through the introduction of an on-line platform aimed at boosting the development of EOSC skills and competences. EOSC-synergy complements on-going activities in EOSC-hub and other related projects liaising national bodies and infrastructures with other upcoming governance, data and national coordination projects.

CPSforTM - Cyber-physical systems (CPS), Internet of Things (IoT), wireless sensor networks, Smart monitoring, tele-medicine, COVID-19

Kyberneticko-fyzický systém pre epidemiologický tele-monitoring a tele-medicínu pre pacientov s COVID 19

Duration: 1. 1. 2021 - 31. 12. 2022
Program: Medziakademická dohoda (MAD)
Project leader: Ing. Budinská Ivana PhD.
Annotation:The aim of the project is research and development of Cyber-Physical System (CPS) for tele-monitoring and tele-medicine in hospital care during pandemic situations, including the following embedded systems: mechanical system, hardware, software, sensors, IoT devices, communication system and web-based user interface containing the necessary functions for work in the hospital environment. This complex system will be controlled via the Internet. The cyber-physical system for epidemiological monitoring in hospital care will be built by a specialized modular Robot-Assistant, and IoT devices together with a system for a seamless communication and aggregation of sensory and multimodal data for tele-medicine and tele-monitoring of the patients with Covid 19. Thus, the physical contact between the medical staff with patients will be limited. Moreover, the risk of repeated daily exposure to corona virus for physicians during each manipulation will be greatly reduced. The project will boost up the cooperation of two academic institutions with complementary research activities. Participating PhD students and young researchers will gain experience from international cooperation, from which they will benefit in their further professional career.

Development of machine learning models for high-performance computing

Návrh modelov strojového učenia pre vysoko-výkonné počítanie.

Duration: 1. 4. 2020 - 31. 12. 2022
Program: Medziakademická dohoda (MAD)
Project leader: doc. Ing. Hluchý Ladislav CSc.
Annotation:Nowadays, Machine learning (ML) is important and relevant trend in the development of modern computer science. ML is not new concept it was part of the field of artificial intelligence for a long time. The main idea focuses on building algorithms that can learn on their own, but recent advances in computing systems and the availability of big data create impressive progress from checkers solving program to self-moved cars and programs that recognize single face from thousands of photos. One of the applications of ML is to improve the efficiency of computations by tuning parameters of environment. Here we deal with complex computational program as with black box, which dynamic depends on parameters. In this project we propose to investigate process of auto tuning of parallel program on heterogeneous multiprocessor system. This problem is important and has some features, that makes it hard for solving. First of all, data and environment can influence performance of program in crucial way. Also, in some setups algorithm can control scheduling making strategy time-dependant. Secondly, an important feature is the presence of competing users. It is necessary to ensure fair and equal access to resources. Because each user is a rational agent and seeks to increase his or her share of the resource at the expense of others, an unbalanced distribution algorithm can shift the system to an inefficient equilibrium. This is problem of game theory. There are notable connections of ML methods with game theory. For example we can mention GANs (generative Adversarial Nets) corresponding to a two-player game, SVM (support vector machine) connected with zero-sum two-player game, and others. This project proposes the following approaches to address the problem. The first is model-free reinforcement machine learning, which is trying to find the best possible control strategy. Secondly, it is the application of the game-theoretic approach, in which the formalized user behavior is involved in the system. Each user has their own learning algorithm and is a rational agent who wants to get some of the computing resource and maximize its utility function. The project continues previous collaboration between institutes in the design of adaptive programming methods for high-performance computing in heterogeneous multiprocessor environments. Project objectives: The purpose of the project is to analyze and synthesize new machine learning algorithms in concurrent high-performance computing. Existing machine learning algorithms are usually heuristic solutions, the quality of which depends on the data and training parameters. One of the promising areas of better understanding of the work is the application of game-theoretical approaches to modeling the learning process of competing algorithms. The results of the game analysis will help to describe the optimal behavior of users at the point of equilibrium and to calculate the characteristics of schedulers and translation algorithms.

FIRElinks - Fire in Earth System: Science & Society (FIRElinks)

Požiar v systéme Zeme: veda a spoločnosť

Duration: 1. 6. 2019 - 1. 6. 2023
Program: COST
Project leader: RNDr. Glasa Ján CSc.

SOON - Social Network of Machines (SOON)

Sociálna sieť strojov

Duration: 1. 3. 2019 - 30. 4. 2022
Program: ERANET
Project leader: Ing. Balogh Zoltán PhD.
Annotation:This project proposes to investigate the impact of the use of autonomous social agents to optimise manufacturing process in the framework of Industry 4.0. "Social" means that cyber-physical entities will act autonomously in order to optimize an industrial process following behaviour models inspired by human social networks. Currently, in Industry 4.0, smart entities do exist. However, intelligence is localised and intelligent heterogeneous entities cannot communicate together even inside the same shop-floor. Our motivation comes from the observation that, if we want to create a real Internet of Everything that brings together processes, data, things, and people, all these entities have to be connected and follow a shared, easy to understand paradigm. In this project, we propose a holistic multi-agent paradigm that encompasses machines and humans. The presence of human operators is therefore crucial both to teach to and to learn from software agents, via deep learning and data mining algorithms. Agents will take decisions merging and analysing big and heterogeneous data produced by sensors (vibration, temperature, etc.), automation and information systems (such as enterprise resource planning and manufacturing execution system), and humans in real-time. The design and evaluation of the SOON system will be performed through predictive maintenance scenarios in collaboration with three different industrial companies (in Slovakia, Spain and Switzerland). Such collaboration will enable the project consortium to assess the concrete improvement on specific industrial processes. As application scenario, the project will focus on the predictive maintenance tasks. We believe that the arrival of Industry 4.0 revolution combined with recent improvements in machine learning, and the application of autonomous multi-agent architecture can finally bring disruptive innovation in industrial process optimization and modelling.

WISE-ACT - Wider Impacts and Scenario Evaluation of Autonomous and Connected Transport (WISE-ACT)

Širšie dopady a vyhodnocovanie scenárov nasadenia autonómnej a prepojenej dopravy

Duration: 1. 3. 2019 - 31. 3. 2022
Program: COST
Project leader: RNDr. Glasa Ján CSc.

National Projects

Algorithm of collective intelligence: Interdisciplinary study of swarming behaviour in bats

Algoritmus kolektívnej inteligencie: Interdisciplinárne štúdium swarmového správania netopierov.

Duration: 1. 8. 2018 - 31. 7. 2022
Program: APVV
Project leader: Ing. Zelenka Ján PhD.
Annotation:Various algorithms of artificial intelligence inspired by real biological mechanisms are successfully applied in military and civil sector. In this proposed project, cooperative research of four scientific institutions having different basis, methodology and the object of study (biology, computer science and technology) focuses on interdisciplinary study of social self-organizational behaviour of tree-dwelling bats with the aim to develop new meta-heuristic method capable of space exploration. The project has a great potential to bring new scientific knowledge about mechanisms of collective intelligence in social structures of biological organisms, which will contribute to the field of theoretical biology, behavioural nad evolutionary ecology. Additionally, it will contribute to the field of artificial intelligence as it focuses on swarming behaviour of individuals/agents with higher nervous activity and well developed cognitive skills. The model organisms (bats) on which this project is focused are using advanced biological mechanism capable of state space exploration and at the same time preventing group disintegration. This specific characteristic has a great potential for development of new biologically inspired algorithms and methods applicable e.g. in movement coordination of unmanned aerial vehicles. Distinguished contribution of the project is R&D of ultra-light sensors applicable not only in biological research. The project originality is in interconnection of different research areas. Therefore it is important to review it interdisciplinary.
Project web page:http://netopiere.sk/skybat/

ARIEN - Analysis of environmental influences on power industry equipment by the methods of artificial intelligence and cloud computing (ARIEN|

Analýza vplyvu prostredia na zariadenia energetického priemyslu metódami umelej inteligencie a cloudového počítania

Duration: 1. 7. 2021 - 31. 12. 2023
Program: APVV
Project leader: Ing. Kvassay Marcel PhD.
Annotation:The main goal of project ARIEN is to develop & verify a new innovative method for the estimation of local pollution deposition for power industry needs based on the data collected by the SHMU Institute. It will rely on advanced techniques of artificial intelligence, cloud computing and data integration and was not yet applied in Slovak power industry. It will help to prevent outages on 110 kV and 400 kV overhead lines caused by flashovers. The results obtained by this method will help to optimize the design of insulators and improve the quality of maintenance during operation. Artificial intelligence is the logical choice, because so far we have not been able, due to their complexity, to describe the processes of pollution dispersion, deposition and chemical interaction adequately by standard physical and chemical models. Techniques of artificial intelligence and neural networks can find new relationships between the input and the output data. Neural networks, however, require large training sets. Therefore, the second project goal is to develop a method for obtaining a sufficiently large set of synthetic data for their training and testing. The third project goal is to research & develop a new method for measuring the insulator pollution by the analysis of their acoustic emissions captured by contactless detectors placed in their vicinity. Here we rely on the experience of the project partner VUJE with measurement systems installed on overhead lines and able to work during their normal operation. These goals reflect current priorities of European research as well as pressing needs of Slovak power industry: in 2015, the standard STN 33 0405 regulating the assessment of high-voltage insulator pollution deposition and their required cleaning frequency, was withdrawn. The new methodology will help replace the tedious & demanding process mandated by this standard by a new one relying on readily available information and the latest computing techniques.

EWA - Early Warning of Alzheimer

Early Warning of Alzheimer

Duration: 1. 9. 2020 - 31. 8. 2023
Program: Štrukturálne fondy EÚ Bratislavský kraj
Project leader: Ing. Rusko Milan PhD.

Electron beam lithography of nanometer structures for 2D materials on the base of metal sulfides

Elektrónová litografia nanometrových štruktúr pre 2D materiály na báze sulfidov kovov

Duration: 1. 1. 2018 - 31. 12. 2021
Program: VEGA
Project leader: Ing. Mgr. Andok Robert PhD.
Annotation:Motivation of this project comes out the research of new 2D materials such as metal sulfides and from the present state of development of e-beam lithography (EBL) as one of the alternative methods of 2D structures preparation in electronics. This project is focused on obtaining new scientific results in the patterning of nano-scale structures (10-100nm) in electron resists. Attention will be paid to the research of the influence of the electron lithography processes on the resulting nanometric patterns patterned in electron resits in term of resolution, dimensions accuracy and edge roughness of the structures in the resist. We will study the parameters influencing the profile of the nanometric patterns in polymeric resists. The simulations of lithographic parameters of electron resists on thin semiconductor layers, nitride membranes and layers with 2D materials, are an important part of this project. Based on the simulations of lithographic parameters, we will prepare nanometric structures on these materials.

ICONTROL - Intelligent Cloud Workflow Management for Dynamic Metric- Optimized Application Deployment (ICONTROL)

Inteligentné riadenie tokov práce v cloude pre dynamické a metrikami optimalizované nasadzovanie aplikácií

Duration: 1. 7. 2021 - 31. 12. 2023
Program: APVV
Project leader: doc. Ing. Hluchý Ladislav CSc.
Annotation:ICONTROL is a platform for dynamic and intelligent function-based workflow application deployment and run-time redeployment in hybrid edge cloud computing environments. ICONTROL will automatically construct and deploy complex workflow-based cloud applications in an end-to-end manner. ICONTROL will use semantic information on available cloud-based application functionalities to automatically generate complex workflows of functions, which will be executed across edge and cloud resources. It will help users to automate function selection, configuration and deployment, dealing with run-time failure or performance issues through automated redeployment. To remove obstacles for non-cloud expert users to use complex function-based workflow applications, ICONTROL will support a separation of tasks among 3 categories of specialists. Cloud application developers will develop application and backend functions, semantically annotate and describe them. AI-powering algorithms will assist domain application experts to create application workflows using functionalities provided by the cloud application developers, with semi-automated workflow construction techniques by leveraging semantic descriptions. Application users will be able to use their application workflows to instantiate, deploy and run their applications using the automation capabilities of the proposed system in complex hybrid edge and cloud environments, not burdened by complexities related to workflows selection, application deployment, fault resolution, resource elasticity provision, to name just a few. ICONTROL will significantly improve composability and adaptability of workflow-based applications spanning the entire edge-cloud continuum (from remote wireless IoT sensors, through personal devices, to large computing centers), by creating a semantically-enriched, platform-agnostic, and secured FaaS workflow development and execution platform, with automatic infrastructure resource management and provisioning.

New Methods and Approaches for Distributed Scalable Computing

Nové metódy a prístupy pre distribuované škálované počítanie

Duration: 1. 1. 2020 - 31. 12. 2022
Program: VEGA
Project leader: doc. Ing. Hluchý Ladislav CSc.
Annotation:Nowadays, there is a growing number of heterogeneous data from distributed sources, which brings great challenges to extract valuable knowledge from them. Current solutions are computationally-intensive modeling and simulations in a variety of science, industry and commercial areas through machine learning, neural networks, and deep learning. Due to the extremely multi-faceted dynamic data it is highly necessary to design novel methodologies, robust methods and approaches for scalable analytics in conjunction with scalable data collection, processing and management. High-performance platforms also need to be upgraded with the latest cloud technology knowledge for flexible management of large-scale systems. The project will also include research and development of appropriate tools and services for distributed and scalable information processing with the support for high-performance platforms. The proposed project builds on the results achieved in the VEGA 2/0167/16 project and in the four H2020 projects.

ORBIS - Ontology representation for security of information systems

Ontologická reprezentácia pre bezpečnosť informačných systémov

Duration: 1. 7. 2020 - 30. 6. 2024
Program: APVV
Project leader: Ing. Budinská Ivana PhD.
Annotation:According to the action plan of the government of the Slovak Republic, one of the priorities is to “support research in cyber security”. This project proposal responds to the action plan's requirements by setting research objectives and goals in the field of cybersecurity. The project proposal focuses on new approaches in sharing knowledge about security incidents and indicators. There are several initiatives that seek to formalize and standardize security incidents descriptions. However, it is neither realistic nor desirable to assume that there will be one common standard for describing security incidents. The solution can be offered by creating a core ontology, which semantically integrates various approaches to describing threats and attacks, thus enabling the integration of several standards and knowledge repositories. The tools for representation and processing of ontologies will be used for this purpose. It will be a combination of procedures for collecting data from network communication, creating the necessary ontologies eg. semiautomatic extraction of ontologies from text in natural language – ontology learning, use of selected advanced methods of ontological knowledge representation, e.g. contextualized representation or methodology of ontological metamodeling. A significant shift will be that the ontology of the model can be represented by a graph. The results can be expected to contribute to more efficient sharing, representation, storage, and use of cybersecurity knowledge.

Computer modelling of fire dynamics and effects

Počítačové modelovanie dynamiky požiaru a jeho dôsledkov

Duration: 1. 1. 2020 - 31. 12. 2022
Program: VEGA
Project leader: RNDr. Glasa Ján CSc.
Annotation:Research in the suggested project is oriented on formulation of new scientific knowledge in the field of computer modelling of fire dynamics and effects. Huge societal damages on property, environment, destroyed buildings and infrastructure and threat for human lives and health caused by fires create acute need for such research of all-society importance. The research will utilize the authors‘ results and experiences on experimental testing of fires and investigation of applicability of models and program systems capable to simulate complex processes related to fire. The research will focus on computer simulation of fire course and consequences of chosen fire scenarios and its efficient parallel realization on high-performance computer systems.

SenArtI - Processing of sensor data via Artificial Intelligence methods.

Spracovanie údajov zo senzorov prostriedkami umelej inteligencie

Duration: 1. 1. 2019 - 31. 12. 2022
Program: VEGA
Project leader: Ing. Malík Peter PhD.
Annotation:The project will propose new methods and algorithms for processing of multi-sensor data for tasks of object diagnostics, areas of interest evaluation, secure communication and simplification of the new intelligent models creation. The research will focus on advanced methods of artificial intelligence with emphasis on deep learning. Artificial Intelligence algorithms show significantly better results than traditional methods, and an example of this is the tremendous progress of refining semantic image segmentation through deep learning over the last five years. Modern, low-cost, miniature electro-mechanical structures allow simple integration and group deployment of multiple sensors, resulting in the production of a huge amount of multi-sensor data that can not be manually processed. The project output will include new models of accurate semantic image segmentation, precision object modeling, multiple sensor collaboration models, and secure data transfer, especially for the Internet of Things and Industry 4.0.

Projects total: 17