Sebastian Matysik
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Sebastian Matysik has a total of 22 co-authors. Here are the top co-authors by number of shared publications:
- Paweł Karol Frankowski
Maritime University of Szczecin
6 shared publications Joanna Wisniewska
4 shared publications- Joanna Wiśniewska
Institute of Management, University of Szczecin
4 shared publications Joanna Wiśniewska
3 shared publicationsPaweł Karol Frankowski
3 shared publications- Ryszard Daniel ZięTek
Department of Robotics and Control, Faculty of Mechatronics and Electrical Engineering, Maritime University of Szczecin, Willowa 2, 71-650 Szczecin, Poland
2 shared publications
Latest Publications
- Open Access
Feature-Level Decoupling of Reinforced-Concrete Parameters from Sparse MFL Data via Three-Layer Sensing
This methodological paper presents a data-efficient, explainable artificial intelligence (XAI) pipeline for non-destructive testing (NDT) of reinforced concrete (RC) under sparse-data conditions, addressing three closely connected challenges: cross-parameter interference in magnetic flux leakage (MFL) signals, scarce training data, and black-box opacity. Data scarcity is a recognized constraint across AI-driven NDT and is structural in reinforced concrete (costly labeling, absent benchmarks, small-data learning missing from the reviewed electromagnetic literature). This scarcity has two consequences: black-box models can generalize poorly on sparse data, and where errors are likely, engineers hesitate to trust or assume liability for opaque predictions. We present a glass-box framework identifying rebar diameter (D), cover thickness (h), and steel class through feature-level decoupling grounded in measurement physics. Transparency and accuracy are treated as equally vital NDT dimensions. Amplitude–Correlation–Offset (ACO) decomposition of a three-layer magnetoresistive transducer isolates parameter-specific attributes: Bx offset classifies steel class (95% confidence), genetic-algorithm fitting localizes peaks to 0.1 mm for D (10–12 mm), and inter-layer amplitude ratios estimate h (20–60 mm). Rough-set analysis quantifies rule confidence and uncertainty zones. Proof-of-concept validation (20 specimens and 400 measurements) shows that a few reference configurations suffice for interpretable identification, enabling explainable artificial intelligence (XAI) in sparse-data NDT and lowering the technology-adoption barrier in safety-critical inspection.
Open AccessFeature-Level Decoupling of Reinforced-Concrete Parameters from Sparse MFL Data via Three-Layer Sensing
This methodological paper presents a data-efficient, explainable artificial intelligence (XAI) pipeline for non-destructive testing (NDT) of reinforced concrete (RC) under sparse-data conditions, addressing three closely connected challenges: cross-parameter interference in magnetic flux leakage (MFL) signals, scarce training data, and black-box opacity. Data scarcity is a recognized constraint across AI-driven NDT and is structural in reinforced concrete (costly labeling, absent benchmarks, small-data learning missing from the reviewed electromagnetic literature). This scarcity has two consequences: black-box models can generalize poorly on sparse data, and where errors are likely, engineers hesitate to trust or assume liability for opaque predictions. We present a glass-box framework identifying rebar diameter (D), cover thickness (h), and steel class through feature-level decoupling grounded in measurement physics. Transparency and accuracy are treated as equally vital NDT dimensions. Amplitude–Correlation–Offset (ACO) decomposition of a three-layer magnetoresistive transducer isolates parameter-specific attributes: Bx offset classifies steel class (95% confidence), genetic-algorithm fitting localizes peaks to 0.1 mm for D (10–12 mm), and inter-layer amplitude ratios estimate h (20–60 mm). Rough-set analysis quantifies rule confidence and uncertainty zones. Proof-of-concept validation (20 specimens and 400 measurements) shows that a few reference configurations suffice for interpretable identification, enabling explainable artificial intelligence (XAI) in sparse-data NDT and lowering the technology-adoption barrier in safety-critical inspection.
- Open Access
SnowBallSLR: saturation-aware and reproducible citation searching for systematic reviews
- Open Access
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 143326-143347
Mapping Scientific Knowledge in Systematic Literature Reviews Through Embeddings and Bibliometric Techniques
- 1 January 2026
- Published by Institute of Electrical and Electronics Engineers (IEEE) in IEEE Access
- Vol. 14, 143326-143347
Open AccessMapping Scientific Knowledge in Systematic Literature Reviews Through Embeddings and Bibliometric Techniques
- Open Access
- 18 November 2025
- Published by MDPI AG in Sustainability
- Vol. 17 (22) , 10307
Integrating Electromagnetic NDT and IoT for Enhanced Structural Health Monitoring of Corrosion in Reinforced Concrete as a Key to Sustainable Smart Cities
The paper addresses a critical gap in early-stage corrosion detection in reinforced concrete, a leading cause of structural failures with significant impacts on humans, the economy, and the environment. It presents the M5 (Magnetic Force-Induced Vibration Evaluation) method, an innovative Structural Health Monitoring (SHM) approach that avoids damping in concrete by using electromagnetic excitation and transferring rebar vibrations through magnetic coupling over the sample. By inducing and analyzing natural vibrations directly in reinforcement, M5 enables sensitive, non-destructive evaluation (NDE) of corrosion before deterioration occurs. The study follows a systematic literature review based on PRISMA standards and utilizes EmbedSLR v1.0 free software. The methodology combines NDE with IoT deployment using Low-Power Wide Area Networks (LPWANs) and advanced machine learning (ARA) to detect frequency changes caused by corrosion, ensuring continuous monitoring. Findings suggest that M5 has the potential to enhance sustainable asset management by extending infrastructure lifespan, optimizing maintenance, and reducing waste. Its practical implications are significant for urban planners and engineers aiming to align infrastructure management with smart city strategies. The originality of this work lies in integrating electromagnetic NDT with IoT and data-driven decision-making, offering new insights at the intersection of engineering and sustainable smart city management.
- 18 November 2025
- Published by MDPI AG in Sustainability
- Vol. 17 (22) , 10307
Open AccessIntegrating Electromagnetic NDT and IoT for Enhanced Structural Health Monitoring of Corrosion in Reinforced Concrete as a Key to Sustainable Smart Cities
The paper addresses a critical gap in early-stage corrosion detection in reinforced concrete, a leading cause of structural failures with significant impacts on humans, the economy, and the environment. It presents the M5 (Magnetic Force-Induced Vibration Evaluation) method, an innovative Structural Health Monitoring (SHM) approach that avoids damping in concrete by using electromagnetic excitation and transferring rebar vibrations through magnetic coupling over the sample. By inducing and analyzing natural vibrations directly in reinforcement, M5 enables sensitive, non-destructive evaluation (NDE) of corrosion before deterioration occurs. The study follows a systematic literature review based on PRISMA standards and utilizes EmbedSLR v1.0 free software. The methodology combines NDE with IoT deployment using Low-Power Wide Area Networks (LPWANs) and advanced machine learning (ARA) to detect frequency changes caused by corrosion, ensuring continuous monitoring. Findings suggest that M5 has the potential to enhance sustainable asset management by extending infrastructure lifespan, optimizing maintenance, and reducing waste. Its practical implications are significant for urban planners and engineers aiming to align infrastructure management with smart city strategies. The originality of this work lies in integrating electromagnetic NDT with IoT and data-driven decision-making, offering new insights at the intersection of engineering and sustainable smart city management.
- Open Access
- 17 November 2025
- Published by International Conference on Information Systems Development
A Comparative Analysis of Embedding Models and Traditional Methods for Publication Selection in Systematic Literature Reviews - A Case Study in Gamification Marketing
This study aimed to compare the effectiveness of the deep ANN embedding technique with traditional article selection methods in systematic literature reviews. The embedding model utilizes a natural language problem description to find semantically similar publications. Consequently, this technique is accessible to users without experience in data exploration. Traditional methods are represented by precise keyword queries in Scopus and Excel-based selection. Keywords used in these methods are extracted from the description by the GPT-4o model with a temperature set to zero, ensuring repeatability. The obtained results were evaluated using bibliometric metrics, which facilitate the assessment of similarities among filtered publications and enhance understanding of their connections. The findings demonstrated the superiority of the embedding model, achieving higher thematic coherence and more shared references and keywords. This approach improves the identification of relevant publications and significantly contributes to automating systematic literature reviews, which is desired in many scientific disciplines.
- 17 November 2025
- Published by International Conference on Information Systems Development
Open AccessA Comparative Analysis of Embedding Models and Traditional Methods for Publication Selection in Systematic Literature Reviews - A Case Study in Gamification Marketing
This study aimed to compare the effectiveness of the deep ANN embedding technique with traditional article selection methods in systematic literature reviews. The embedding model utilizes a natural language problem description to find semantically similar publications. Consequently, this technique is accessible to users without experience in data exploration. Traditional methods are represented by precise keyword queries in Scopus and Excel-based selection. Keywords used in these methods are extracted from the description by the GPT-4o model with a temperature set to zero, ensuring repeatability. The obtained results were evaluated using bibliometric metrics, which facilitate the assessment of similarities among filtered publications and enhance understanding of their connections. The findings demonstrated the superiority of the embedding model, achieving higher thematic coherence and more shared references and keywords. This approach improves the identification of relevant publications and significantly contributes to automating systematic literature reviews, which is desired in many scientific disciplines.
- Open Access
- 4 November 2025
- Published by University of Szczecin
- p. 344-365
ARTIFICIAL INTELLIGENCE FOR PERSONALIZED RANKINGS OF ARTICLES – NOVEL EMBEDDING METHOD SUPPORTING DECISION-MAKING IN SYSTEMATIC LITERATURE REVIEW
Purpose: This work presents a new Artificial Intelligence (AI) method to support decision-making in selecting optimal articles from the user's perspective. The technique leads to selecting high-quality works that fit the researcher's needs by enabling hierarchizing scientific publications based on their similarity to a predefined researcher problem. This article is the first in a planned series on AI-powered systematic literature reviews. Need for the Study: The rapid development of AI over the last few years has caused the number of publications to grow rapidly and their quality to decrease. Developing a filtering method to select adequate papers is imperative in an AI-dominated environment where classic methods are no longer effective. Methodology: The embedding model transforms research queries and publication metadata into numerical vector representations. The cosine similarity metric ranks publications based on their semantic closeness to the research query. Next, bibliometric analyses validate the effectiveness of the selection process. Findings: The proposed AI-powered selection method significantly improves the thematic consistency of selected publications compared to entire datasets downloaded from the Scopus database. Selected articles exhibit higher bibliometric similarity in shared references, keywords, and co-citations. The approach efficiently filters large datasets while maintaining relevance, providing researchers with a refined subset of highly pertinent publications. Practical Implications: The proposed method can improve systematic literature reviews by automating the selection of relevant publications, reducing manual workload, and ensuring thematic accuracy. The findings suggest that AI-driven text embedding techniques can enhance the efficiency and reliability of literature filtering, benefiting researchers across various disciplines.
- 4 November 2025
- Published by University of Szczecin
- p. 344-365
Open AccessARTIFICIAL INTELLIGENCE FOR PERSONALIZED RANKINGS OF ARTICLES – NOVEL EMBEDDING METHOD SUPPORTING DECISION-MAKING IN SYSTEMATIC LITERATURE REVIEW
Purpose: This work presents a new Artificial Intelligence (AI) method to support decision-making in selecting optimal articles from the user's perspective. The technique leads to selecting high-quality works that fit the researcher's needs by enabling hierarchizing scientific publications based on their similarity to a predefined researcher problem. This article is the first in a planned series on AI-powered systematic literature reviews. Need for the Study: The rapid development of AI over the last few years has caused the number of publications to grow rapidly and their quality to decrease. Developing a filtering method to select adequate papers is imperative in an AI-dominated environment where classic methods are no longer effective. Methodology: The embedding model transforms research queries and publication metadata into numerical vector representations. The cosine similarity metric ranks publications based on their semantic closeness to the research query. Next, bibliometric analyses validate the effectiveness of the selection process. Findings: The proposed AI-powered selection method significantly improves the thematic consistency of selected publications compared to entire datasets downloaded from the Scopus database. Selected articles exhibit higher bibliometric similarity in shared references, keywords, and co-citations. The approach efficiently filters large datasets while maintaining relevance, providing researchers with a refined subset of highly pertinent publications. Practical Implications: The proposed method can improve systematic literature reviews by automating the selection of relevant publications, reducing manual workload, and ensuring thematic accuracy. The findings suggest that AI-driven text embedding techniques can enhance the efficiency and reliability of literature filtering, benefiting researchers across various disciplines.
- Open Access
- 4 November 2025
- Published by University of Szczecin
- p. 366-388
EFFICIENT AI-POWERED DECISION-MAKING IN SYSTEMATIC LITERATURE REVIEWS USING MULTI-EMBEDDING MODELS
Purpose: This paper presents a new Artificial Intelligence (AI) multi-model approach for selecting publications that fit researcher needs well. It is the second work in a series of AI-powered systematic literature reviews focusing on replacing the human research team with a multi-model framework. Need for the study: The expected exponential growth of scientific literature generated fully or partially by AI demands advanced artificial intelligence tools to streamline systematic literature reviews. Such a method should filter databases according to the researcher's needs, minimizing the risk of overlooking critical works. Methodology: The technology behind word embedding models (WEMs) powers various AI systems, including ChatGPT. Eight models of this kind were evaluated with bibliometric criteria to select and utilize the four top-performing ones in article selection. Each model independently selects publications that best fit the problem description. Papers selected by the larger number of models are considered more relevant. Findings: Different WEMs are trained on different data, have different architectures, and recognize different publications as the most relevant. Usually, articles selected by more models show higher thematic coherence; however, the number of models in the framework cannot be too large. The process is analogous to a systematic literature review conducted by many researchers with all the advantages and disadvantages of their parallel work. Practical Implications: Our method is less biased and more accurate than a multiple-person systematic literature review (SLR). It allows for a fast examination of the problem from many perspectives and has the same implications as SLR.
- 4 November 2025
- Published by University of Szczecin
- p. 366-388
Open AccessEFFICIENT AI-POWERED DECISION-MAKING IN SYSTEMATIC LITERATURE REVIEWS USING MULTI-EMBEDDING MODELS
Purpose: This paper presents a new Artificial Intelligence (AI) multi-model approach for selecting publications that fit researcher needs well. It is the second work in a series of AI-powered systematic literature reviews focusing on replacing the human research team with a multi-model framework. Need for the study: The expected exponential growth of scientific literature generated fully or partially by AI demands advanced artificial intelligence tools to streamline systematic literature reviews. Such a method should filter databases according to the researcher's needs, minimizing the risk of overlooking critical works. Methodology: The technology behind word embedding models (WEMs) powers various AI systems, including ChatGPT. Eight models of this kind were evaluated with bibliometric criteria to select and utilize the four top-performing ones in article selection. Each model independently selects publications that best fit the problem description. Papers selected by the larger number of models are considered more relevant. Findings: Different WEMs are trained on different data, have different architectures, and recognize different publications as the most relevant. Usually, articles selected by more models show higher thematic coherence; however, the number of models in the framework cannot be too large. The process is analogous to a systematic literature review conducted by many researchers with all the advantages and disadvantages of their parallel work. Practical Implications: Our method is less biased and more accurate than a multiple-person systematic literature review (SLR). It allows for a fast examination of the problem from many perspectives and has the same implications as SLR.
- 1 November 2025
- Published by ISMA SYC INT in EUROPEAN RESEARCH STUDIES JOURNAL
- Vol. XXVIII (4) , 242-260
Structural Health Monitoring of Corrosion in Reinforced Concrete: A Key Component for Smart Cities
- 1 November 2025
- Published by ISMA SYC INT in EUROPEAN RESEARCH STUDIES JOURNAL
- Vol. XXVIII (4) , 242-260
Structural Health Monitoring of Corrosion in Reinforced Concrete: A Key Component for Smart Cities