Artificial Intelligence and Intelligent Systems

In our fast-changing world, one of the technologies sticks out. The technology that will change our daily routines, change the way we learn, how we work, how we buy goods, how we spend our free time, how we commute, the way we find friends and communicate, and the way we reach out to the stars. The technology that will change everything. Artificial intelligence. We, a group of fifty researchers, innovators, and educators, are proposing to continue to explore new ways, methods, and algorithms that build up artificial intelligence. In the past, we have already invented techniques and tools that have contributed to its growth and that the scientists have recognized as its key components. We created ways to make artificial intelligence and its models easier to understand, ways to explain its predictions, and means to integrate large volumes of heterogeneous data. We embedded these methods into tools and offered them to the community of scientists, researchers, and educators. We enriched local industry with our knowledge of data-driven management and production, devised applications to help some of the most prominent Slovenian industries to grow and keep the comparative advantage, and collaborated with startups to define their cutting-edge technologies. We have collaborated with hospitals and clinical centers to improve the health care system and clinical treatments, devised means for the assisted living of the elderly, and contributed to advancements in other research fields, such as molecular biology and ecology. We will continue to develop, innovate, disseminate, collaborate, and educate. We are here proposing a continuation of our Programme. In developing new methods and tools, we still focus on explainability, interpretation, use of domain knowledge, visual presentation of data, models, and discovered patterns, engagement of domain experts, and data fusion. These areas have formed the foundation of our research in the past and have recently become the hottest research topics with the omnipresence of AI and the success of deep neural networks. We propose to structure our work to address the development of AI methods, applications, tools, and education. We will innovate data mining and visualization, deep learning, recommender systems, argument-based learning, qualitative modeling, evolutionary computation and heuristic search, and agent-based learning. In all these approaches, we will specifically address explanation, interpretation, and, where applicable, fairness of derived models. We will collaborate with industry, healthcare providers, and the public sector to apply AI and deliver cutting-edge technology to our community. We will continue to make some of the world's best software tools for AI and continue to impress our peers with their utility and robustness. And finally, we plan further to develop our approach to conceptual training of AI, and promote its deliverance to schools, universities, industry, and community.

We propose to research several crucial and timely topics in artificial intelligence that are very important due to the impact and applicability of this fast-developing area of science. The programme will significantly contribute to the development of new methods that address these topics, create practical and useful software environments that support both research and utility of these methods, and confirm the applicability of our new scientific results in practice. We will continue to address the problems of explainability and interpretation of AI models and the inclusion of humans in the model development loop. The creation of transparent and human-readable models that can include available domain knowledge has been a determining trait of our programme since its beginnings. We have already contributed world-renowned techniques to address these problems in feature scoring, model explanation, and data fusion approaches. The recent emergence and popularity of deep models, the availability of large data sets, and surge of AI applications in all, including sensitive, areas of life have exposed the need for explainability, interpretability and faithfulness. We are pleased that we have a head start in this area of research and can continue to make significant contributions in the future. Our proposed research addresses disciplines beyond computer science. The AI methods, including machine learning, data mining, exploration of domain knowledge, data fusion, search and optimization, and design of intelligent systems, became vital for the infrastructure of other scientific disciplines, both in natural and social sciences. Today, AI impacts all the fields where data abounds, and with digitalization, AI is becoming omnipresent. The programme addresses these other disciplines through applied research and collaborations with partner institutions. Our mission is to develop AI approaches to contribute to the development of science, and our duty to use it wisely and educate about its reach and impact.

Digitalization and its upgrade with AI have recently been set as top priorities in the technological development of the EU and Slovenia. Public administration in Slovenia has recently crafted a "National Program to Promote the Development and Use of AI until 2025" (NpAI), which is now being signed by the Slovene Government. Our programme proposes actions that will contribute to education, training, transfer of knowledge, and implementation of AI techniques both in industries and the public sector. Our proposed programme is well-aligned with recent developmental initiatives, including NpAI. We have and will continue to design AI approaches that can advance our industrial partners. We will continue to collaborate with major industries and SMEs in the country, including industries in steel and appliance manufacturing, pharmacy, retail, and automatization (see Section 32). We have creative collaboration with selected SMEs that can place our products and services on the market. And we will strive to innovate in developing end-user AI software, including our flagship data mining toolbox Orange. An important part of our programme addresses emerging needs in healthcare and assisted living. AI is a crucial component of emerging healthcare tools, thus helping healthcare systems cope with the aging population and increasing chronic patients. It also demonstrated that it could assist in emergencies such as the recent COVID-19 pandemic. Personal health systems are an affordable way of bringing quality healthcare to a broad population, improving equality of access to health services. Assisted living systems powered by AI can help seniors age in place, preserving their quality of life and contributing to the sustainability of social care systems. The research programme - in collaboration with domain experts from various fields, such as molecular biologists and clinical experts - will contribute by developing medical data analysis methods, devices and methods for assisted living, tools for personalized medicine, and by searching for novel biomarkers and diagnostic/screening methods based on emerging sensor technologies (e.g. eye-tracking, wearables, etc.). We build on our long-term collaboration with medical professionals; some of these collaborations even predate this programme (est. in 1999). AI can enhance cultural experiences. While this is not a major direction of our research, the programme has contributed to providing access to digitized and physical cultural heritage for citizens and tourists in the past. We will be on the lookout for opportunities to do so again. We also closely collaborate with several public administration institutions, including the Ministry of Public Administration and Ministry of Agriculture, Forestry and Food. This collaboration aims to advance their digital infrastructures and seek opportunities where AI implementation would improve the quality of their services and support decision-making. We offer hands-on courses designed for public administration employees and have carried out about ten workshops that showcased the use of machine learning. We aim to continue with such training and carry on with pilot projects that demonstrate the use of AI in the public sector. At present, our collaboration with the Ministry of Public Administration in the development of semantic analysis of documents has been recognized as a flagship R&D project in the public sector that uses AI. Many members of the programme group are university teachers. With the importance of the AI methods and their great potential socioeconomic influence, it is essential to teach these methods and their practical use to new generations of students. Our strong collaborations with foreign partners will additionally enrich socioeconomic impact. The group established relations with numerous prestigious foreign institutions and brought several foreign scientists on shorter and longer visits to Slovenia. We also plan to involve them in teaching at the University of Ljubljana and Jožef Stefan Institute.