Selected work

Representative Papers

Coordinated incentives in AI-generated misinformation governance

Aug 07, 2026

This study addresses the escalating spread of misinformation fueled by the proliferation of AI-generated content, which undermines information credibility and societal trust. To analyze this challenge, the authors construct an evolutionary game-theoretic model involving three key stakeholders—government regulators, AI firms, and users—and incorporate heterogeneous reward-punishment mechanisms. By employing replicator dynamics, they examine strategic interactions among these agents. The findings reveal that neither regulatory oversight nor market-based incentives alone suffice to curb misinformation effectively. Instead, an evolutionarily stable equilibrium favoring truthful content production emerges only when coordinated incentives align such that regulatory intensity, reputational penalties for noncompliance, and user-level rewards jointly exceed critical thresholds. This work thus provides a theoretical foundation and identifies essential parameter conditions for the collaborative governance of AI-driven misinformation.

0 citationsRead paper

Scholarly Production and Public Health Determinants in Context of Funding: The Case of IoMT Research:

Jun 21, 2026

This study addresses the unclear scale of funding in Internet of Medical Things (IoMT) research and its impact on scholarly output and national public health indicators. Integrating bibliometric analysis, topic modeling, and triangulation methods, the work systematically compares publication trends, thematic distributions, and associations with health outcomes between funded and unfunded IoMT studies. It reveals, for the first time, a significant positive correlation between IoMT funding intensity and national social determinants of health, while demonstrating that funded research is more concentrated on cutting-edge areas such as artificial intelligence. The findings suggest that research funding not only accelerates scholarly productivity in the IoMT domain but may also enhance healthcare quality by fostering technological innovation.

0 citationsRead paper

Splitting User Stories Into Tasks with AI -- A Foe or an Ally?

May 08, 2026

This study addresses the time-consuming and error-prone nature of decomposing user stories into actionable tasks in agile development, highlighting the need for effective support mechanisms. The authors present the first empirical evaluation—through controlled experiments and developer surveys—of a generative AI tool (GitLab Duo) for task decomposition, alongside a proposed human-AI collaborative workflow. Findings indicate that AI assistance yields more granular and comprehensive task lists, though manual filtering is required to remove irrelevant suggestions. Developers consistently favored a hybrid approach that integrates AI-generated outputs with traditional methods, as it enhances planning efficiency while preserving accuracy and contextual relevance.

0 citationsRead paper

Framework for identifying the equivalence between Nature-Inspired Metaheuristics

Mar 30, 2026

This study addresses growing concerns that many nature-inspired metaheuristic algorithms are merely variants of existing methods, lacking objective criteria to assess equivalence. To resolve this, the work formally defines strong equivalence between metaheuristics and introduces a general discrimination framework based on cosine similarity of phenotypic and genotypic feature vectors. This approach overcomes the limitations of subjective evaluation and provides a theoretical foundation for assessing algorithmic novelty. Extensive experiments demonstrate that, under realistic computational constraints, mainstream algorithms rarely achieve high similarity thresholds, indicating that most newly proposed methods are not trivial replications. These findings validate the effectiveness and practical utility of the proposed framework in distinguishing genuinely novel metaheuristics from superficial variants.

0 citationsRead paper

Towards Data-driven Nitrogen Estimation in Wheat Fields using Multispectral Images

Feb 24, 2026

Accurate targeted nitrogen fertilization for wheat is challenged by multiple confounding factors—including crop type, growth stage, soil conditions, and weather—making precise estimation of nitrogen demand difficult. To address this, this work proposes TerrAI, a novel deep neural network–based approach that, for the first time, integrates multispectral remote sensing data with spatiotemporal modeling to explicitly capture the spatial and temporal heterogeneity of nitrogen dynamics across fields. By fusing spatiotemporal features from multispectral imagery, TerrAI enables data-driven, high-precision prediction of wheat nitrogen requirements. Experimental results on real-world remote sensing datasets demonstrate that the method significantly improves nitrogen estimation accuracy, offering an effective and scalable solution for intelligent fertilization decision-making in precision agriculture.

0 citationsRead paper

Recent publications

Latest Papers

Coordinated incentives in AI-generated misinformation governance

Aug 07, 2026

This study addresses the escalating spread of misinformation fueled by the proliferation of AI-generated content, which undermines information credibility and societal trust. To analyze this challenge, the authors construct an evolutionary game-theoretic model involving three key stakeholders—government regulators, AI firms, and users—and incorporate heterogeneous reward-punishment mechanisms. By employing replicator dynamics, they examine strategic interactions among these agents. The findings reveal that neither regulatory oversight nor market-based incentives alone suffice to curb misinformation effectively. Instead, an evolutionarily stable equilibrium favoring truthful content production emerges only when coordinated incentives align such that regulatory intensity, reputational penalties for noncompliance, and user-level rewards jointly exceed critical thresholds. This work thus provides a theoretical foundation and identifies essential parameter conditions for the collaborative governance of AI-driven misinformation.

0 citationsRead paper

Scholarly Production and Public Health Determinants in Context of Funding: The Case of IoMT Research:

Jun 21, 2026

This study addresses the unclear scale of funding in Internet of Medical Things (IoMT) research and its impact on scholarly output and national public health indicators. Integrating bibliometric analysis, topic modeling, and triangulation methods, the work systematically compares publication trends, thematic distributions, and associations with health outcomes between funded and unfunded IoMT studies. It reveals, for the first time, a significant positive correlation between IoMT funding intensity and national social determinants of health, while demonstrating that funded research is more concentrated on cutting-edge areas such as artificial intelligence. The findings suggest that research funding not only accelerates scholarly productivity in the IoMT domain but may also enhance healthcare quality by fostering technological innovation.

0 citationsRead paper

Splitting User Stories Into Tasks with AI -- A Foe or an Ally?

May 08, 2026

This study addresses the time-consuming and error-prone nature of decomposing user stories into actionable tasks in agile development, highlighting the need for effective support mechanisms. The authors present the first empirical evaluation—through controlled experiments and developer surveys—of a generative AI tool (GitLab Duo) for task decomposition, alongside a proposed human-AI collaborative workflow. Findings indicate that AI assistance yields more granular and comprehensive task lists, though manual filtering is required to remove irrelevant suggestions. Developers consistently favored a hybrid approach that integrates AI-generated outputs with traditional methods, as it enhances planning efficiency while preserving accuracy and contextual relevance.

0 citationsRead paper

Framework for identifying the equivalence between Nature-Inspired Metaheuristics

Mar 30, 2026

This study addresses growing concerns that many nature-inspired metaheuristic algorithms are merely variants of existing methods, lacking objective criteria to assess equivalence. To resolve this, the work formally defines strong equivalence between metaheuristics and introduces a general discrimination framework based on cosine similarity of phenotypic and genotypic feature vectors. This approach overcomes the limitations of subjective evaluation and provides a theoretical foundation for assessing algorithmic novelty. Extensive experiments demonstrate that, under realistic computational constraints, mainstream algorithms rarely achieve high similarity thresholds, indicating that most newly proposed methods are not trivial replications. These findings validate the effectiveness and practical utility of the proposed framework in distinguishing genuinely novel metaheuristics from superficial variants.

0 citationsRead paper

Towards Data-driven Nitrogen Estimation in Wheat Fields using Multispectral Images

Feb 24, 2026

Accurate targeted nitrogen fertilization for wheat is challenged by multiple confounding factors—including crop type, growth stage, soil conditions, and weather—making precise estimation of nitrogen demand difficult. To address this, this work proposes TerrAI, a novel deep neural network–based approach that, for the first time, integrates multispectral remote sensing data with spatiotemporal modeling to explicitly capture the spatial and temporal heterogeneity of nitrogen dynamics across fields. By fusing spatiotemporal features from multispectral imagery, TerrAI enables data-driven, high-precision prediction of wheat nitrogen requirements. Experimental results on real-world remote sensing datasets demonstrate that the method significantly improves nitrogen estimation accuracy, offering an effective and scalable solution for intelligent fertilization decision-making in precision agriculture.

0 citationsRead paper