IIPhDW 2025

A summary of the 2025 International Interdisciplinary PhD Workshop, a conference combined with the fourth edition of workshops on innovative technologies and applications of artificial intelligence, was presented. The event was organized at the Lublin University of Technology and Management (WSEI) and took place from September 11–13, 2025. The program included a plenary session, thematic sessions, a poster session, and networking events. The opening ceremony, with an introductory lecture, was led by Professor Tomasz Rymarczyk.

The event’s thematic scope encompassed technical computer science, electrical engineering, sensor systems, and artificial intelligence, with a particular emphasis on process and medical tomography engineering, machine learning, cyberphysical systems, and industrial and medical applications. The conference was designed with PhD students and the scientific community in mind. The multi-year and international nature of the series, which in previous editions has been held in Germany, the Czech Republic, Slovakia, and Poland, was emphasized. Emphasis was placed on combining fundamental research with implementations in industrial and service areas.

On the first day, after the opening, papers were presented on new methods of image reconstruction and signal analysis in impedance tomography and ultrasound. Papers included papers on physics-informed neural networks, multi-head architectures, differential deep networks, and tomographic image generation. Participants focused on the methodology of combining modeling with machine learning and aspects of measurement data quality.

The poster session on the first day covered a wide range of topics. Presentations included solutions for crystallization control using hybrid tomography and reinforcement learning, reconstruction methods using residual networks, systems for non-invasive monitoring of the lower urinary tract, and work on pattern recognition and natural language processing in smart services. Projects also included UWB-based object localization and time-of-flight measurements, system performance testing using stress tools, and the integration of reconstruction and ML algorithms in process monitoring. The posters also covered cloud resource management, cybersecurity issues, and the application of deep learning methods in spectral analysis.

The second day began with an opening lecture, followed by thematic sessions. Research results were presented on the effectiveness of holography and mixed reality in process tomography surveillance, personalized modeling and prediction of bladder behavioral changes in patients, and non-invasive glucose measurement using impedance methods. Work was presented on radar positioning systems in indoor spaces and an analysis of scheduling heuristics in container environments for machine learning applications.

The event highlighted strategic directions for the development of artificial intelligence technologies and their impact on innovation in industry and medicine. Attention was paid to the complementarity of deterministic and machine learning methods, the importance of measurement quality, and the need to integrate solutions into closed-loop systems. An overview of trends and timelines for the development of language models was presented, as well as directions for practical implementations in the areas of tomography, automation, the Internet of Things, and agent systems.

The key results and organizational conclusions were summarized by the scientific committee and the session management teams. The high level of expertise presented, the clear representation of projects with significant implementation potential, and the need for further collaboration on reference data packages and metric standardization in tomography, signal and image analysis were emphasized. The event was deemed a success both scientifically and organizationally, with its format ensuring a balance between the presentation of results and discussion of key topics for the development of research and innovation.