IKTHM EN

AI, HPC and BigData

Artificial intelligence is increasingly entering people's daily lives and is also having an increasing impact on the economy. In Slovenia we have a well-developed research sphere in this field, but it is necessary to take a step forward in the use of AI in the economy, egg. in supporting business and production processes and upgrading products and services. Slovenia is internationally recognized in the field of AI research and development, but it's that its introduction into the business processes of the economy and the state administration is still slow. It brings immense opportunities for optimization and development in all areas of the government, society and economy. The use of AI technologies enables new, quick insights into events, greatly shortens development cycles, enables machine quality control, even where conditions are unfavorable for humans, and above all should contribute to improving the quality of life in general.

 In the focus area AI, HPC & BigData, we will strive for a faster transfer of 40 years of experience and research results into the work processes of the economy in particular. As analyzes show in the near future AI will have a decisive impact on the competitiveness and productivity of companies. We will advocate the implementation of a national program for the development and introduction of artificial intelligence. We will conduct webinars, events and trainings for the purpose of raising awareness and building trust. With the SRIP domain domains, we will identify areas for new potential joint products. Over the last decade, a number of companies have emerged which have artificial intelligence either as a core business or use it in day-to-day operations. We will build on the starting points of the current product directions of HPC and BigData, as the focus, using artificial intelligence methods, has mainly shifted from the infrastructure (HPC and BigData) on the wider scope of their use.

For faster transfer of artificial intelligence into practice the initiative AI4SI (AI for Slovenia) works within the focus area AI, HPC & BigData. It represents is a connecting link between solution providers and researchers in the field of artificial intelligence and companies that want to use artificial intelligence in their work and also includes important stakeholders who cannot be members of the SRIP but are crucial for the development of the field.

The prospects for artificial intelligence are very high, as its ability to detect and predict patterns from vast amounts of data far exceeds human capabilities, it represents tremendous opportunities in forecasting, decision support, automation of operations, product optimization, new business models, targeted services, user experience and even in the field of environmental protection and public health. Artificial intelligence is already changing the way we communicate with computer systems. At Accenture, they analyzed 12 developed economies and found that artificial intelligence has the potential to double their annual economic growth rates by 2035. The consulting firm PwC predicts that artificial intelligence will increase global GDP by 14 percent or 15 percent in 2030, $ 7 trillion - more than the GDP of China and India combined.

HPC oz. high-performance computing accelerates the development of the industry, encourages it to innovate products and services faster and more efficiently through the use of super-fast computers, modeling tools and simulations, and provides additional impetus to industries that offer services and products. The field of "Big Data" offers approaches and technologies for management and data analytics and decision support. The data it handles can be of various sizes, from small to very large, for which we need a special high-performance cloud infrastructure, HPC or HTC. Data sources can be static and recorded in databases, or they can be obtained in real time from sensors or online sources and processed on an ongoing basis.

Here we can emphasize the competitive advantages of Slovenia in: (1) modeling and simulation of complex engineering problems (eg modeling of drugs, seismic analysis); (2) developing new customized system programs for orchestrating applications, making optimal use of computational resources, achieving reliability, high levels of availability, service-level contracts, etc .; (3) cloud computing technologies that are particularly suitable for large distributed data storage applications and for time-critical applications. Slovenia can also boast of an excellent ICT infrastructure (eg ARNES, ARCTUR, inclusion in the PRACE program, EGI, etc.), which also includes a developed network ICT infrastructure (eg optical network along motorways in Slovenia), which is the basis for the development of new highly distributed smart applications.

In Slovenia in the field of "big data" ("Big Data") there is a lot of existing knowledge and engineering skills, and in some areas (eg text technology, image analysis and computer vision, data merging, teaching tools) we are in the world top. The high level of knowledge originates from the already mentioned foundation in the field of artificial intelligence. Due to its connection to artificial intelligence, the Slovenian School of Machine Learning  especially promotes the understanding of data and the development of intuitive user interfaces, which can represent a significant competitive advantage in the field of business intelligence and decision support technologies.

 

In order to focus product directions, as well as to respond to the current situation and orientations of national and EU policies and measures and expected development incentives, the partnership will focus on the following product directions and connect with other focus and domain areas to achieve the set goals:

PS1 AI solutions for the field of language technologies

If we want to increase the user experience when using digital products and services, it is crucial to improve the interaction with smart devices using multimodal user interfaces. Speech is one of the most natural ways to interact. For successful development and use of speech-supported applications, it is necessary to provide efficient and high-quality components of the speech dialogue system, ie the performance of automatic speech recognition, understandable and natural-sounding speech synthesis that allows automatic speech signal generation based on input text. In doing so, we quickly encounter a gap between technologically "supported" and "unsupported" languages, which is deepening in the world; that is, between languages ​​where users can use advanced speech and language technology solutions and those that are more or less "useless" for these purposes. As part of the project, which includes the largest educational and research organizations in Slovenia, a central infrastructure for language resources and technologies for the Slovenian language will be established. The project, co-financed by the Republic of Slovenia and the European Union from the European Regional Development Fund, lasts from May 2020 to August 2022. Based on the identified needs for missing computer products and services in the field of language technologies, the project aims to establish infrastructure for language resources and technologies. open access to research results that will be made public and available under a Creative Commons (CC) license. In this way, researchers and scientific research organizations, start-ups and other companies, as well as the wider international community, will be enabled to upgrade organized storage, maintenance and distribution of resources and technologies. The latter also ensures more equal cooperation of the Slovenian research area with European initiatives for the exchange of digital resources and tools. Today’s large language models of the BERT and GPT-3 type from large text collections capture language knowledge and apply it to a variety of tasks. They can also do this successfully for multiple languages ​​at the same time and currently achieve the best results e.g. in summarizing texts, classifying sentiment, and answering questions. This development did not bypass Slovene either.

PS2 AI solutions for smart cities and communities

AI in smart cities will play an important role in smarter urbanization with the aim of sustainable growth and improving economic and living conditions. Big data and artificial intelligence (AI) are interconnected. AI can efficiently process large amounts of big data and create forecasts and cost-effective solutions. It will help address challenges such as traffic congestion, security controls, car park management, crime prevention, accident detection, autonomous aircraft, resource management (energy, water, rubbish,), digital twin cities, improving management and planning, decarbonisation, optimization mobility, public safety. We will focus primarily on the development and implementation of solutions in the fields of health, resource supply, mobility and public safety, as well as providing the building blocks of AI technologies. The solutions will connect via an urban open data platform. The development will be based on the processing and use of city and community data, the development of an AI widget for an open urban data platform and the definition of data standards.

PS3 AI solutions for the transition to a circular economy

The transition to a large-scale circular economy will only be possible through the use of digital technology and the introduction of new, sustainable business models. In-depth insights gained from big data analysis from securely connected IoT systems, automation and robotics, artificial intelligence technologies will be the heart of this transformation and a great opportunity to optimize the use of resources and processes. The vast amounts of data captured by devices and other systems represent an unimaginable potential for the transition to a circular economy. UI can improve and accelerate the development of new products, components, and materials through iterative machine learning procedures that enable rapid prototyping and testing. With the help of UI, we can search for sustainable circular business models through simulations and check scenarios of how the offer of services and the lease of devices, instead of buying, can increase the competitiveness of the company. By combining data collected in real time with historical data, we can predict the use of funds, propose optimal prices, anticipate demand, initiate preventive maintenance, and manage inventories wisely.

PS4 AI solutions for sustainable food production and processing

Ensuring sustainable safe food from the field to the table will depend on data exchange. Artificial intelligence technologies make it possible to increase agricultural productivity with smart agricultural equipment and machinery, flexible energy use for cooling, supply chain optimization, path optimization, quality control, process improvement, production automation and reduction of food waste. By recognizing the state of development of pests, it enables the optimization of spray consumption. Weather forecasting models make it possible to plan agricultural tasks.

PS5 Geospatial AI (GeoAI)

Artificial intelligence (AI) has been at the forefront of most technological change in the field of spatial informatics. Using intelligent algorithms, sorting and recognizing spatial data data from a variety of sources, and using predictive analysis, machine learning based on geographic components are opportunities to use in many sectors. Machine vision, neural networks, and deep learning are transforming current systems and applications and enabling new and fast models of use. Using dedicated building blocks that we will develop for an open data platform, we will use artificial intelligence technologies to process data and better understand the world around us.

PS6 AI solutions to enhance security using UI

Security analysts investigate threats by finding patterns, formulating hypotheses, inferring, and sharing their findings with other companies. These are labor-intensive steps that require not only time but also years of training and expertise. As operational technology (OT) continues to integrate with the corporate network, and as threat actors continue to advance in their methods of attack, the gap in cybersecurity in the OT world is becoming increasingly apparent. Rapid advances in the Internet and communications have further led to a rapid increase in the size of networks and the amount of relevant data. As a result, many new attacks are created, which pose an additional security challenge for networks, as fast and accurate intrusion detection is becoming more and more demanding. AI systems used in Slovenian companies to ensure security are currently, without exception, imported foreign products. The logic underlying the decision-making mechanisms between what is a threat and what is not is invisible, unknown and conceived abroad. The wish is to use Slovenian knowledge and experience to create an AI system that would be able to at least partially replace the currently used foreign products.

Despite the remarkable efforts of researchers, these systems still face challenges in improving the speed and accuracy of incident detection while reducing the rate of false alarms and in detecting new intrusions. Recently, deep learning (GU) -based systems for effective security incident detection have been introduced as potential solutions. Members have devised a solution that will be able to detect several types of cyber incidents. The prototype is designed in such a way that the ability to detect other types of incidents can be added to it later. The prototype solution captures data from the network, processes it in real time and, in the event of a detected incident, alerts them. Based on the detected incident, the security engineer takes appropriate action. The main goals of the solution are to increase the speed of cyber incident detection, increase the accuracy of cyber incident detection, reduce the level of false alarms, the ability to detect new cyber incidents and provide Slovenia with its own high-performance cyber security technologies.

PS7 AI solutions for health and medicine

Awareness of capabilities, timely ethical principles and regulations, transparency and thus the trust of all users are key to greater use of artificial intelligence technologies in healthcare and medicine. Advances in artificial intelligence (AI) could prevent many misdiagnosis and treatment of symptoms instead of the causes of the disease. Huge amounts of data, including laboratory test results, ECGs, and X-rays, that are collected and stored in electronic health records, allow the use of UI and high-performance data-based medicine. UI is already changing the way doctors and researchers approach solving clinical problems. Some algorithms may clash with physicians, but they still cannot be fully integrated into everyday medical practice on their own. However, they are increasingly in the role of assistant to the doctor. Although these algorithms have a significant impact on the effectiveness of treatment and medical interventions, there are still some regulatory and ethical issues to which clear answers need to be sought in advance. Healthcare is about life, and in such cases extreme care is needed when introducing new technologies. UI in medicine and healthcare involves the use of machine learning algorithms to mimic human cognitive function in the analysis, interpretation, and understanding of complex medical and health data. The increasing availability of health data and the rapid advancement of analytical techniques allow the use of UI usually in early detection of disease, diagnosis, treatment of patients and prediction of outcome, but also allows the use of UI in other related areas such as. finding safe and effective drug combinations. However, AI also plays an important role in tackling the COVID-19 pandemic. Members will develop solutions and services for the effective and safe use of medicines, individualized dosing of medicines, detection of diseases such as dementia, in the field of radiotherapy and mixed reality. We will participate in defining common data spaces in the field of health data and thus create conditions for trustworthy, ethical and safe use of data.

 PS8 Development and implementation of AI in business processes

 The volume of available data has increased exponentially in recent years, researchers have developed more sophisticated machine and deep learning algorithms, and the computational power and storage capacity have been steadily increasing. The convergence of these trends encourages rapid technological progress and abrupt changes in the economy and society. All of these technologies and algorithms are available to all organizations. UI technologies today do not require high investment, they are even available as a rental service. Analytical houses predict that artificial intelligence will have a decisive impact on the competitiveness and productivity of companies in the near future, and successful companies are already taking advantage of its advantages. Successful companies in this field focus mainly on energy and utilization of UI to create added value. People and processes are key here, and technology is just a tool. In Slovenia, we have a well-developed research sphere in the field of artificial intelligence, but it is necessary to take a step forward in the use of artificial intelligence in the economy in supporting business and production processes and in upgrading the products and services themselves. We will focus on the development and implementation of AI in the business processes of organizations. We will develop predictive analytics solutions and services for sales and marketing (personalization and forecasting), decision support systems, resource optimization, business process optimization, business process automation, controlling, price optimization, anomaly detection and consumer sentiment detection.

 We will participate in defining common data spaces, data models and create conditions for trustworthy, ethical and secure use of data.

 

 

  The AI4SI Group (AI for Slovenia) will strive for a faster transfer of research results into the work processes of the economy in particular, as analyzes show that in the near future IM will have a decisive impact on the competitiveness and productivity of companies. The group will include representatives of the economy, state administration and representatives of research and development organizations active in the field of artificial intelligence and other stakeholders. 
You are invited to the AI4SI initiative website, https://ai4si.gzs.si/. Also follow the LinkedIn profile.