
Big data and the ever-growing need for their fast and efficient processing, aimed at obtaining information used in the automation of business processes, improving communication, enhancing efficiency and success in practically every segment of human activities, as well as better understanding of nature, humans, and society, are some of the main characteristics of computer technology in the 21st century. These characteristics largely determine contemporary technological trends: machine learning, artificial intelligence, and the processing of large amounts of data (signals) have emerged as some of the most attractive fields of scientific research in modern times. Directly or indirectly, these scientific fields dictate success in overcoming problems and the efficiency of solutions in other areas—energy, transportation, communications, security, (bio)medicine and epidemiology, climate science, ecology, genetics, microbiology, physics, as well as in social sciences—economics, sociology, etc.
The processing of data, or signals on graphs, is one of the responses to the challenges posed by their increasing volume and the need to consider the internal, structural characteristics of these data, as well as their often hidden and not always obvious interdependencies. Signal processing on graphs has also emerged as a response to the need for better modeling of structural and topological characteristics of data acquisition systems (for example, sensor networks), to be utilized as effectively as possible in the process of analysis and information extraction, and in the process of machine learning (for example, learning the topology of a graph).
The proposed project aims to enable the development of fundamental theory in the field of signal processing on graphs and the development of new algorithms, techniques, and methods that allow for the effective application of theoretical concepts in practice. Sophisticated techniques for data analysis and processing can be directly applied to specific data—multimedia content (digital images, sound, video, text, and other data), biomedical signals (including epidemiological modeling and analyses), data obtained from sensor measurements, aiming at knowledge extraction, noise removal, quality improvement, data reconstruction, so one of the project’s goals is the successful experimental verification of the developed concepts.