Eva Volna
Current Organizations / Institutes
The organizations/institutes from the scholar's latest article
Previous Organizations / Institutes
The organizations / institutes from the scholar's other articles in the past ten years
Research Subjects
Research Interests
Eva Volna has a total of 67 co-authors. Here are the top co-authors by number of shared publications:
- Martin Kotyrba
Department of Informatics and Computers, Faculty of Science, University of Ostrava, 30. dubna 22, 701 03 Ostrava, Czech Republic
46 shared publications Martin Kotyrba
University of Ostrava
16 shared publications- Hashim Habiballa
Department of Informatics and Computers, Faculty of Science, University of Ostrava, 30. Dubna 22, 70103 Ostrava, Czech Republic
12 shared publications - Robert Jarusek
Department of Informatics and Computers, University of Ostrava, 30 dubna 22, Ostrava, Czech Republic
7 shared publications Vaclav Kocian
University of Ostrava, Czech Republic
6 shared publicationsMichal Janosek
Centre of Excellence IT4Innovations Division of the University of Ostrava
5 shared publications
Latest Publications
- Open Access
- 29 September 2026
- Published by MDPI AG in Machine Learning and Knowledge Extraction
- Vol. 8 (10) , 303
Do Large Language Models Prefer Their Developer’s Home Country? A US–China Audit of Scores Under National Attribution
We examine how developer origin, citizenship-perspective instructions, and English versus Mandarin presentation shape sentiment scores from large language models (LLMs) for nationality-matched actions. Across two experiments using the same eleven-model panel, we collected 4,065,600 response records: 2,032,800 for the main corpus and 2,032,800 for the earlier corpus. Each corpus contains 26,400 bilingual stimulus rows, comprising 13,200 nationality-matched pairs across 22 event categories, two intended polarities, and three temporal frames, evaluated under seven conditions. Inference about developer origin uses ten models from the United States (US) and China and a nine-developer sensitivity analysis. In the main experiment, the English baseline developer-origin gap is score points (exact model-label permutation ; developer-level ). Citizenship-perspective instructions produce language-averaged attribution changes of points for the US persona and for the China persona. All ten models shift in the requested direction in both English persona conditions; nine do so in each Mandarin condition. Mandarin changes absolute scoring, especially for positive-design items, while its average factorial effect on the paired attribution contrast is small and uncertain. The temporal origin interaction is also uncertain ( points; ), despite polarity-dependent temporal patterns. Non-numeric returns account for of main-experiment responses and occur only among hosted models. The experiments converge on substantial average persona steering and modest, uncertain baseline origin differences, with model-specific exceptions. Corpus content, persona wording, and collection period vary together, and reuse of the panel adds no independent origin units. The findings identify explicit perspective instructions as a consequential part of the scoring protocol within this fixed panel. The audit concerns nationality-matched generic scenarios and does not establish equivalence between origin groups or the absence of broader geopolitical bias.- 29 September 2026
- Published by MDPI AG in Machine Learning and Knowledge Extraction
- Vol. 8 (10) , 303
Open AccessDo Large Language Models Prefer Their Developer’s Home Country? A US–China Audit of Scores Under National Attribution
We examine how developer origin, citizenship-perspective instructions, and English versus Mandarin presentation shape sentiment scores from large language models (LLMs) for nationality-matched actions. Across two experiments using the same eleven-model panel, we collected 4,065,600 response records: 2,032,800 for the main corpus and 2,032,800 for the earlier corpus. Each corpus contains 26,400 bilingual stimulus rows, comprising 13,200 nationality-matched pairs across 22 event categories, two intended polarities, and three temporal frames, evaluated under seven conditions. Inference about developer origin uses ten models from the United States (US) and China and a nine-developer sensitivity analysis. In the main experiment, the English baseline developer-origin gap is score points (exact model-label permutation ; developer-level ). Citizenship-perspective instructions produce language-averaged attribution changes of points for the US persona and for the China persona. All ten models shift in the requested direction in both English persona conditions; nine do so in each Mandarin condition. Mandarin changes absolute scoring, especially for positive-design items, while its average factorial effect on the paired attribution contrast is small and uncertain. The temporal origin interaction is also uncertain ( points; ), despite polarity-dependent temporal patterns. Non-numeric returns account for of main-experiment responses and occur only among hosted models. The experiments converge on substantial average persona steering and modest, uncertain baseline origin differences, with model-specific exceptions. Corpus content, persona wording, and collection period vary together, and reuse of the panel adds no independent origin units. The findings identify explicit perspective instructions as a consequential part of the scoring protocol within this fixed panel. The audit concerns nationality-matched generic scenarios and does not establish equivalence between origin groups or the absence of broader geopolitical bias. - Open Access
- 24 September 2026
- Published by Prague University of Economics and Business in Acta Informatica Pragensia
Enterprise Conversational AI with Retrieval-Augmented Generation: A CRM Case Study
A peer-reviewed journal of applied informatics that covers a broad range of topics, especially dealing with the social and business aspects of informatics.
- 24 September 2026
- Published by Prague University of Economics and Business in Acta Informatica Pragensia
Open AccessEnterprise Conversational AI with Retrieval-Augmented Generation: A CRM Case Study
A peer-reviewed journal of applied informatics that covers a broad range of topics, especially dealing with the social and business aspects of informatics.
- Open Access
- 24 June 2026
- Published by Elsevier in SSRN Electronic Journal
Grata Cascade: Hash-based One-Shot Key Agreement via Statistical Drift as a Candidate for Hybrid Post-Quantum Key Establishment
Modern hybrid key establishment combines Diffie–Hellman and Module-LWE assumptions to achieve post-quantum security. We present Grata Cascade, a one-shot key agreement primitive based exclusively on a third, structurally independent assumption: SHA-256 preimage resistance combined with statistical drift of a publicly reconstructable shared state. The protocol converges in a median of approximately 49 rounds, derives a 32-byte shared key, and provides a cumulative preimage barrier of λ = 120 bits per session. Our contributions are threefold. First, we provide a structural security argument based on a generalized birthday asymmetry amplified across rounds. Second, we experimentally evaluate the protocol over 5 × 10⁴ convergence runs, obtaining 100% convergence and 99.992% key agreement, with an upper 95% Clopper–Pearson confidence bound on the failure rate of 2.05 × 10⁻⁴. Randomness evaluation using NIST SP 800-22 passes all 12 tests over 1000 derived keys. We further evaluate five classes of passive adversaries on 300 transcripts each, observing no successful key recoveries. Third, we compare Grata Cascade with established one-shot key agreement mechanisms, including ECDH, ML-KEM, X-Wing hybrid, and Merkle puzzles. The protocol incorporates three complementary structural defenses: a cumulative 120-bit preimage barrier achieved through multiple hash-filtering stages, a large indistinguishable candidate set exceeding the preimage barrier, and oracle-free AES-CBC-based key derivation. Finally, we discuss extending Grata Cascade toward a Continuous Key Agreement (CKA) primitive as an open research direction. The present work establishes Grata Cascade as a candidate for a third, structurally independent layer in hybrid post-quantum key establishment.
- 24 June 2026
- Published by Elsevier in SSRN Electronic Journal
Open AccessGrata Cascade: Hash-based One-Shot Key Agreement via Statistical Drift as a Candidate for Hybrid Post-Quantum Key Establishment
Modern hybrid key establishment combines Diffie–Hellman and Module-LWE assumptions to achieve post-quantum security. We present Grata Cascade, a one-shot key agreement primitive based exclusively on a third, structurally independent assumption: SHA-256 preimage resistance combined with statistical drift of a publicly reconstructable shared state. The protocol converges in a median of approximately 49 rounds, derives a 32-byte shared key, and provides a cumulative preimage barrier of λ = 120 bits per session. Our contributions are threefold. First, we provide a structural security argument based on a generalized birthday asymmetry amplified across rounds. Second, we experimentally evaluate the protocol over 5 × 10⁴ convergence runs, obtaining 100% convergence and 99.992% key agreement, with an upper 95% Clopper–Pearson confidence bound on the failure rate of 2.05 × 10⁻⁴. Randomness evaluation using NIST SP 800-22 passes all 12 tests over 1000 derived keys. We further evaluate five classes of passive adversaries on 300 transcripts each, observing no successful key recoveries. Third, we compare Grata Cascade with established one-shot key agreement mechanisms, including ECDH, ML-KEM, X-Wing hybrid, and Merkle puzzles. The protocol incorporates three complementary structural defenses: a cumulative 120-bit preimage barrier achieved through multiple hash-filtering stages, a large indistinguishable candidate set exceeding the preimage barrier, and oracle-free AES-CBC-based key derivation. Finally, we discuss extending Grata Cascade toward a Continuous Key Agreement (CKA) primitive as an open research direction. The present work establishes Grata Cascade as a candidate for a third, structurally independent layer in hybrid post-quantum key establishment.
- Open Access
- 21 May 2026
- Published by MDPI AG in Mathematics
- Vol. 14 (10) , 1782
Optimization of Material Permeability Analysis Algorithm for 3D Raster Structures Using Graph-Based and Morphological Approaches
Quantitative characterization of permeability in porous media represents a central problem in filtration theory, geosciences, and materials engineering. Standard numerical approaches, including finite element methods and Lattice Boltzmann simulations, typically require extensive domain-specific expertise together with specialized computational software. This motivates the development of computationally simpler and more accessible geometric approaches applicable directly to binary volumetric data. We introduce a novel algorithmic framework for the analysis of porous structures that reformulates permeability-related characterization in terms of discrete geometry and graph-based computation. The method combines parallel raster-grid and graph representations of a binarized three-dimensional CT image. The principal transport-limiting feature of the pore network, interpreted as the minimal constriction governing connectivity, is identified through iterative morphological dilation coupled with a three-dimensional scanline seed-fill procedure. In addition, a dichotomous bisection strategy is proposed to accelerate the determination of the critical bottleneck scale. The proposed methodology was evaluated on five volumetric datasets of size 100 × 100 × 100 voxels obtained from CT-derived porous structures. Experimental results demonstrate that dilation- and erosion-based formulations yield equivalent estimates of the bottleneck parameter in four of the five investigated samples. Furthermore, incorporation of the bisection optimization reduces computational time in three-dimensional experiments by approximately 50% relative to sequential iteration. The presented approach provides a computationally efficient and fully open-source alternative to conventional physics-based permeability solvers for binary porous media. The resulting bottleneck parameter b should be interpreted as a discrete geometric invariant characterizing the pore-network connectivity and minimal transport cross-section. It is not intended to replace the absolute permeability coefficient K appearing in Darcy’s law, but rather to serve as an independent structural descriptor suitable for comparative and topological analysis of porous systems.
- 21 May 2026
- Published by MDPI AG in Mathematics
- Vol. 14 (10) , 1782
Open AccessOptimization of Material Permeability Analysis Algorithm for 3D Raster Structures Using Graph-Based and Morphological Approaches
Quantitative characterization of permeability in porous media represents a central problem in filtration theory, geosciences, and materials engineering. Standard numerical approaches, including finite element methods and Lattice Boltzmann simulations, typically require extensive domain-specific expertise together with specialized computational software. This motivates the development of computationally simpler and more accessible geometric approaches applicable directly to binary volumetric data. We introduce a novel algorithmic framework for the analysis of porous structures that reformulates permeability-related characterization in terms of discrete geometry and graph-based computation. The method combines parallel raster-grid and graph representations of a binarized three-dimensional CT image. The principal transport-limiting feature of the pore network, interpreted as the minimal constriction governing connectivity, is identified through iterative morphological dilation coupled with a three-dimensional scanline seed-fill procedure. In addition, a dichotomous bisection strategy is proposed to accelerate the determination of the critical bottleneck scale. The proposed methodology was evaluated on five volumetric datasets of size 100 × 100 × 100 voxels obtained from CT-derived porous structures. Experimental results demonstrate that dilation- and erosion-based formulations yield equivalent estimates of the bottleneck parameter in four of the five investigated samples. Furthermore, incorporation of the bisection optimization reduces computational time in three-dimensional experiments by approximately 50% relative to sequential iteration. The presented approach provides a computationally efficient and fully open-source alternative to conventional physics-based permeability solvers for binary porous media. The resulting bottleneck parameter b should be interpreted as a discrete geometric invariant characterizing the pore-network connectivity and minimal transport cross-section. It is not intended to replace the absolute permeability coefficient K appearing in Darcy’s law, but rather to serve as an independent structural descriptor suitable for comparative and topological analysis of porous systems.
- 16 November 2025
- Published by Springer Nature
- p. 322-333
Metabolomic Predictions via SOM: A Cold-Stress Case Study in Arabidopsis thaliana
- 16 November 2025
- Published by Springer Nature
- p. 322-333
Metabolomic Predictions via SOM: A Cold-Stress Case Study in Arabidopsis thaliana
- Open Access
- 9 January 2025
- Published by MDPI AG in Applied Sciences
- Vol. 15 (2) , 581
Artificial Intelligence (ChatGPT) and Bloom’s Taxonomy in Theoretical Computer Science Education
The study focuses on evaluating the performance of AI-based tools, specifically ChatGPT versions 3.5 and 4.0, in comparison to human students in the field of Theoretical Computer Science Education. The experiment aims to assess the capabilities of both AI and human subjects in solving learning tasks based on Bloom’s Taxonomy. The primary objectives of the study are to determine the educational performance of AI and students, identify areas where students may outperform AI in learning tasks, and evaluate the normalized overall results of educational performance with equal assignment weights. The assessment included testing on various types of Bloom’s Learning Task Taxonomy (LTT) based on specific Learning Objectives (LOs). Hypotheses were formulated for quantitative and qualitative analysis to compare the performance of AI and human subjects. Quantitative analysis revealed engaging results regarding the educational performance evaluation with AI-based tools. While some differences were observed in the performance of AI and students, the normalized overall results with equal assignment weights did not show statistically significant differences. The study highlighted the advantages and disadvantages of both humans and AI bots in solving learning tasks. The study concludes that there are areas where students may have an advantage over AI in tasks requiring understanding, evaluation, and creative thinking. Recommendations are provided for educators on utilizing AI-based tools in education, emphasizing the coexistence of AI possibilities with well-designed assignments. The findings suggest that AI has a valuable role in education, and a thorough analysis of teaching approaches and student evaluation is essential in leveraging artificial intelligence effectively.
- 9 January 2025
- Published by MDPI AG in Applied Sciences
- Vol. 15 (2) , 581
Open AccessArtificial Intelligence (ChatGPT) and Bloom’s Taxonomy in Theoretical Computer Science Education
The study focuses on evaluating the performance of AI-based tools, specifically ChatGPT versions 3.5 and 4.0, in comparison to human students in the field of Theoretical Computer Science Education. The experiment aims to assess the capabilities of both AI and human subjects in solving learning tasks based on Bloom’s Taxonomy. The primary objectives of the study are to determine the educational performance of AI and students, identify areas where students may outperform AI in learning tasks, and evaluate the normalized overall results of educational performance with equal assignment weights. The assessment included testing on various types of Bloom’s Learning Task Taxonomy (LTT) based on specific Learning Objectives (LOs). Hypotheses were formulated for quantitative and qualitative analysis to compare the performance of AI and human subjects. Quantitative analysis revealed engaging results regarding the educational performance evaluation with AI-based tools. While some differences were observed in the performance of AI and students, the normalized overall results with equal assignment weights did not show statistically significant differences. The study highlighted the advantages and disadvantages of both humans and AI bots in solving learning tasks. The study concludes that there are areas where students may have an advantage over AI in tasks requiring understanding, evaluation, and creative thinking. Recommendations are provided for educators on utilizing AI-based tools in education, emphasizing the coexistence of AI possibilities with well-designed assignments. The findings suggest that AI has a valuable role in education, and a thorough analysis of teaching approaches and student evaluation is essential in leveraging artificial intelligence effectively.
- 1 January 2024
- Published by AIP Publishing in AIP Conference Proceedings
- Vol. 3094 (1) , 080007
Application to train crystallised intelligence according to the CHC intelligence model
The proposed application is designed to train Crystallised Intelligence according to the CHC Intelligence Model. This application is one of the modules of a res The proposed application is designed to train Crystallised Intelligence according to the CHC Intelligence Model. This application is one of the modules of a res
- 1 January 2024
- Published by AIP Publishing in AIP Conference Proceedings
- Vol. 3094 (1) , 080007
Application to train crystallised intelligence according to the CHC intelligence model
The proposed application is designed to train Crystallised Intelligence according to the CHC Intelligence Model. This application is one of the modules of a res The proposed application is designed to train Crystallised Intelligence according to the CHC Intelligence Model. This application is one of the modules of a res
- Open Access
- 16 October 2023
- Published by Springer Nature in BMC Medical Informatics and Decision Making
- Vol. 23 (1) , 1-14
Proposal of neural network model for neurocognitive rehabilitation and its comparison with fuzzy expert system model
This article focuses on the development of algorithms for a smart neurorehabilitation system, whose core is made up of artificial neural networks. The authors of the article have proposed a completely unique transfer of ACE-R results to the CHC model. This unique approach allows for the saturation of the CHC model domains according to modified ACE-R factor analysis. The outputs of the proposed algorithm thus enable the automatic creation of a personalized and optimized neurorehabilitation plan for individual patients to train their cognitive functions. A set of tasks in 6 levels of difficulty (level 1 to level 6) was designed for each of the nine CHC model domains. For each patient, the results of the ACE-R screening helped deter-mine the specific CHC domains to be rehabilitated, as well as the initial gaming level for rehabilitation in each domain. The proposed artificial neural network algorithm was adapted to real data from 703 patients. Experimental outputs were compared to the outputs of the initially designed fuzzy expert system, which was trained on the same real data, and all outputs from both systems were statistically evaluated against expert conclusions that were available. It is evident from the conducted experimental study that the smart neurorehabilitation system using artificial neural networks achieved significantly better results than the neurorehabilitation system whose core is a fuzzy expert system. Both algorithms are implemented into a comprehensive neurorehabilitation portal (Eddie), which was supported by a research project from the Technology Agency of the Czech Republic.
- 16 October 2023
- Published by Springer Nature in BMC Medical Informatics and Decision Making
- Vol. 23 (1) , 1-14
Open AccessProposal of neural network model for neurocognitive rehabilitation and its comparison with fuzzy expert system model
This article focuses on the development of algorithms for a smart neurorehabilitation system, whose core is made up of artificial neural networks. The authors of the article have proposed a completely unique transfer of ACE-R results to the CHC model. This unique approach allows for the saturation of the CHC model domains according to modified ACE-R factor analysis. The outputs of the proposed algorithm thus enable the automatic creation of a personalized and optimized neurorehabilitation plan for individual patients to train their cognitive functions. A set of tasks in 6 levels of difficulty (level 1 to level 6) was designed for each of the nine CHC model domains. For each patient, the results of the ACE-R screening helped deter-mine the specific CHC domains to be rehabilitated, as well as the initial gaming level for rehabilitation in each domain. The proposed artificial neural network algorithm was adapted to real data from 703 patients. Experimental outputs were compared to the outputs of the initially designed fuzzy expert system, which was trained on the same real data, and all outputs from both systems were statistically evaluated against expert conclusions that were available. It is evident from the conducted experimental study that the smart neurorehabilitation system using artificial neural networks achieved significantly better results than the neurorehabilitation system whose core is a fuzzy expert system. Both algorithms are implemented into a comprehensive neurorehabilitation portal (Eddie), which was supported by a research project from the Technology Agency of the Czech Republic.
- Open Access
- 29 August 2023
- Published by MDPI AG in Applied System Innovation
- Vol. 6 (5) , 77
Improving the Production Efficiency Based on Algorithmization of the Planning Process
Planning and managing the production process are key challenges faced by every manufacturing organization. The main contribution of this article lies in the analysis and design of a planning algorithm that takes into consideration the specifics of this environment. The proposed algorithm encompasses elements of batch production, including a just-in-time approach. The article focuses on scenarios within batch production. Managers of manufacturing and supply companies must ensure smooth fulfillment and uninterrupted production of the agreed-upon quantity of parts. However, this task presents complex challenges. The product portfolio requires meticulous sequencing of production batches, and subsequent parts need to be temporarily stored in their raw state for further processing. Moreover, product variability necessitates frequent adjustments to the production line, resulting in delays. Shortages in manpower additionally place demands on shift organization. The company’s primary objective is to increase production efficiency while simultaneously reducing inventory and minimizing non-standard shift work. The challenge was to reconcile seemingly conflicting company requirements and to concentrate on solutions with swift implementation and minimal costs. Ensuring seamless production operation can be addressed by expanding supporting technologies or by increasing production capacity, such as acquiring an additional production line. However, these options entail costs and do not align with the company’s expectation for immediate impact and cost savings. However, improving production efficiency can also be achieved by altering the approach to production planning, which is the central theme of this article. The key element is ensuring that the customer plan is adhered to while working with a fixed production logic and variable input factors that must account for various non-standard situations.
- 29 August 2023
- Published by MDPI AG in Applied System Innovation
- Vol. 6 (5) , 77
Open AccessImproving the Production Efficiency Based on Algorithmization of the Planning Process
Planning and managing the production process are key challenges faced by every manufacturing organization. The main contribution of this article lies in the analysis and design of a planning algorithm that takes into consideration the specifics of this environment. The proposed algorithm encompasses elements of batch production, including a just-in-time approach. The article focuses on scenarios within batch production. Managers of manufacturing and supply companies must ensure smooth fulfillment and uninterrupted production of the agreed-upon quantity of parts. However, this task presents complex challenges. The product portfolio requires meticulous sequencing of production batches, and subsequent parts need to be temporarily stored in their raw state for further processing. Moreover, product variability necessitates frequent adjustments to the production line, resulting in delays. Shortages in manpower additionally place demands on shift organization. The company’s primary objective is to increase production efficiency while simultaneously reducing inventory and minimizing non-standard shift work. The challenge was to reconcile seemingly conflicting company requirements and to concentrate on solutions with swift implementation and minimal costs. Ensuring seamless production operation can be addressed by expanding supporting technologies or by increasing production capacity, such as acquiring an additional production line. However, these options entail costs and do not align with the company’s expectation for immediate impact and cost savings. However, improving production efficiency can also be achieved by altering the approach to production planning, which is the central theme of this article. The key element is ensuring that the customer plan is adhered to while working with a fixed production logic and variable input factors that must account for various non-standard situations.
- Open Access
- 20 December 2022
- Published by MDPI AG in Mathematics
- Vol. 11 (1) , 7
Expert System for Neurocognitive Rehabilitation Based on the Transfer of the ACE-R to CHC Model Factors
This article focuses on developing an expert system applicable to the area of neurocognitive rehabilitation. The benefit of this interdisciplinary research is to propose an expert system that has been adapted based on real patients’ results from the Addenbrooke’s cognitive examination (ACE-R). One of this research’s main results is a unique proposal to transfer the ACE-R result to the CHC (Cattell–Horn–Carroll) intelligence model. This unique approach enables transforming the CHC model domains according to the modified ACE-R factor analysis, which has never been used before. The expert system inference results allow the automated optimized design of a neurorehabilitation plan to train patients’ cognitive functions according to the CHC model. A set of tasks in 6 difficulty levels (Level 1–Level 6) was proposed for each of the nine CHC model domains. For each patient, the ACE-R results helped determine specific CHC domains to be rehabilitated as well as the starting game level for the rehabilitation within each domain. The proposed expert system has been verified on real data of 705 patients and achieved an average error of 5.94% for all CHC model domains. The proposed system is to be included in the outcomes of the research project of the Technology Agency of the Czech Republic as a verified procedure for healthcare providers.
- 20 December 2022
- Published by MDPI AG in Mathematics
- Vol. 11 (1) , 7
Open AccessExpert System for Neurocognitive Rehabilitation Based on the Transfer of the ACE-R to CHC Model Factors
This article focuses on developing an expert system applicable to the area of neurocognitive rehabilitation. The benefit of this interdisciplinary research is to propose an expert system that has been adapted based on real patients’ results from the Addenbrooke’s cognitive examination (ACE-R). One of this research’s main results is a unique proposal to transfer the ACE-R result to the CHC (Cattell–Horn–Carroll) intelligence model. This unique approach enables transforming the CHC model domains according to the modified ACE-R factor analysis, which has never been used before. The expert system inference results allow the automated optimized design of a neurorehabilitation plan to train patients’ cognitive functions according to the CHC model. A set of tasks in 6 difficulty levels (Level 1–Level 6) was proposed for each of the nine CHC model domains. For each patient, the ACE-R results helped determine specific CHC domains to be rehabilitated as well as the starting game level for the rehabilitation within each domain. The proposed expert system has been verified on real data of 705 patients and achieved an average error of 5.94% for all CHC model domains. The proposed system is to be included in the outcomes of the research project of the Technology Agency of the Czech Republic as a verified procedure for healthcare providers.