Applying conceptual integration theory, the paper analyses the construction of meaning of political parody and political humor in which daily news is explicitly deconstructed. In particular, it focuses on Stephen Colbert’s MAGA Frozen parody and political humor. The paper also analyses how the artistic adaptation of the tune “Let it Go” contributes to highlighting certain aspects of reality, and thus achieving rhetorical goals in political discourse. The aim of the paper is twofold: to examine the creation of humorous meaning in multimodal formats in a conceptual blending that is a result of the incongruity produced in a blended space and to examine the role of the novel inferences and conceptualizations produced in the blended space in achieving a number of rhetorical goals, that is, criticism of the political and social reality. Humorous meaning is created in the emergent structure due to the unusual combination of related elements, which results in incongruity (Coulson 2002; Marín-Arrese 2003; Fujii 2008). The variety of modalities employed in the creation of humor, including different visual cues, helps viewers unpack the blends and successfully understand the criticism of the political reality. Additionally, since blending can serve as a rhetorical tool capable of influencing the audience to alter their perception of reality and take action (Coulson 2006; Coulson and Oakley 2006; Coulson and Pascual 2006; Oakley and Coulson 2008), analyzing the satirical text’s meaning construction through conceptual blending can uncover rhetorical and argumentative objectives within political discourse. In this way, certain viewpoints are promoted, and rhetorical goals are achieved through humor. The interpretation and understanding of humor depend on having general knowledge of the political and social situation, as well as on ideological beliefs and knowledge related to specific events. The viewer must create conceptual associations between input spaces using background knowledge to understand the meaning of the blend.

Ovaj rad razmatra Katulovu poeziju u kontekstu rimske tradicije invektive te istražuje način na koji pjesnik preobražava ovaj javni retorički žanr u sferu osobnog i emocionalnog pjesničkog govora. Analizom odabranih pjesama pokazuje se da Katulova invektiva nije puka osobna uvreda, nego književna strategija koja povezuje intimno iskustvo s društvenom kritikom. U njegovim stihovima vulgarizam, groteska i seksualna optužba postaju instrumenti književnog izraza koji istodobno ponižava protivnika i reflektira granice poezije. Rad naglašava hibridnost Katulove poetike: njegova lirika prelazi u satiru, a satira poprima introspektivne i emotivne dimenzije. U usporedbi s kasnijim satiričarima – Horacijem, Juvenalom i Marcijalom – Katul zauzima prijelomno mjesto jer upravo on oblikuje model pjesnika koji je istodobno intimni ispovjednik i javni prokazivač. Time se potvrđuje polazna teza da Katulova originalnost leži u premještanju invektive iz političkog i retoričkog konteksta u prostor osobne poezije, čime se stvaraju temelji za razvoj rimske satire i hibridne književne tradicije.

In the face of increasing financial uncertainty and market complexity, this study presents a novel risk-aware financial forecasting framework that integrates advanced machine learning techniques with intuitionistic fuzzy multi-criteria decision-making (MCDM). Tailored to the BIST 100 index and validated through a case study of a major defense company in T\"urkiye, the framework fuses structured financial data, unstructured text data, and macroeconomic indicators to enhance predictive accuracy and robustness. It incorporates a hybrid suite of models, including extreme gradient boosting (XGBoost), long short-term memory (LSTM) network, graph neural network (GNN), to deliver probabilistic forecasts with quantified uncertainty. The empirical results demonstrate high forecasting accuracy, with a net profit mean absolute percentage error (MAPE) of 3.03% and narrow 95% confidence intervals for key financial indicators. The risk-aware analysis indicates a favorable risk-return profile, with a Sharpe ratio of 1.25 and a higher Sortino ratio of 1.80, suggesting relatively low downside volatility and robust performance under market fluctuations. Sensitivity analysis shows that the key financial indicator predictions are highly sensitive to variations of inflation, interest rates, sentiment, and exchange rates. Additionally, using an intuitionistic fuzzy MCDM approach, combining entropy weighting, evaluation based on distance from the average solution (EDAS), and the measurement of alternatives and ranking according to compromise solution (MARCOS) methods, the tabular data learning network (TabNet) outperforms the other models and is identified as the most suitable candidate for deployment. Overall, the findings of this work highlight the importance of integrating advanced machine learning, risk quantification, and fuzzy MCDM methodologies in financial forecasting, particularly in emerging markets.

The overview of charophytes in Bosnia and Herzegovina has been updated after 35 years through a revision of existing specimens in the BEOU Charophyte Collection, verification and listing of known vouchers in other herbarium collections, and a detailed review of the available literature covering the period from 1848 to 2024. According to all available data, 18 species and three genera of charophytes are found in 122 sites in Bosnia and Herzegovina. Chara papillosa Kütz. and C. subspinosa Rupr. are newly reported species in Bosnia and Herzegovina. Chara vulgaris L., C. contraria A. Braun ex Kütz., C. globularis Thuill., C. gymnophylla (A. Braun) A. Braun, and C. squamosa Desf. are the most frequently recorded. The Dinaric Mountains mixed forests are the most representative and species-rich ecoregion. Most charophytes were recorded before 1930 and after 1980, with a clear discrepancy between the sites documented in these periods. Bosnia and Herzegovina has the lowest charophyte species richness compared to neighbouring countries. We strongly encourage further revision of available collections and continued systematic field research, which will enable the preparation of the Red List and the formal protection of species and habitats.

The extent to which exogenous sources, including cancer treatment, contribute to somatic evolution in normal tissue remains unclear. Here we used high-depth duplex sequencing1 (more than 30,000× coverage) to analyse 168 cancer-free samples representing 16 organs from 22 patients with metastatic cancer enroled in the PEACE research autopsy study. In every sample, we identified somatic mutations (range 305–2,854 mutations) at low variant allele frequencies (median 0.0000323). We extracted 16 distinct single-base substitution mutational signatures, reflecting processes that have moulded the genomes of normal cells. We identified alcohol-induced mutation acquisition in liver, smoking-induced mutagenesis in lung and cardiac tissue, and multiple treatment-induced processes, which correlated with therapy type and duration. Exogenous sources, including treatment, underpinned, on average, more than 40% of mutations in liver but less than 10% of mutations in brain samples. Finally, we observed tissue-specific selection, with positive selection in tissues such as lung (PTEN and PIK3CA), liver (NF2L2) and spleen (BRAF and NOTCH2), and limited selection in others, such as brain and cardiac tissue. More than 25% of driver mutations in normal tissue exposed to systemic anti-cancer therapy, including in TP53, could be attributed to treatment. Immunotherapy, although not associated with increased mutagenesis, was linked to driver mutations in PPM1D and TP53, illustrating how non-mutagenic treatment can sculpt somatic evolution. Our study reveals the rich tapestry of mutational processes and driver mutations in normal tissue, and the profound effect of lifetime exposures, including cancer treatment, on somatic evolution. High-depth sequencing of non-cancerous tissue from patients with metastatic cancer reveals single-base mutational signatures of alcohol, smoking and cancer treatments, and reveals how exogenous factors, including cancer therapies, affect somatic cell evolution.

This study analyzes the relationship between leadership style and innovative work behavior of employees in companies in Bosnia and Herzegovina. The empirical research was conducted on a sample of 116 respondents employed in various sectors within the territory of the Federation of Bosnia and Herzegovina. Data were collected using a written survey technique through an adapted questionnaire based on the Multifactor Leadership Questionnaire (MLQ) and the scale for measuring innovative work behavior (De Jong & Den Hartog, 2010). Descriptive statistics, reliability analysis (Cronbach’s Alpha), and correlation analysis were applied for data processing. The results indicate that the transactional leadership style is more pronounced compared to the laissez-faire leadership style. The levels of innovative work behavior of employees showed the highest values in the dimensions of idea exploration and idea generation, while idea promotion and idea implementation were less represented. The correlation analysis revealed that there are statistically significant, but generally low to moderate positive relationships between leadership styles and innovative work behavior. The strongest interrelationships were observed among the different dimensions of innovative work behavior, confirming that innovative activities constitute an interconnected process. Based on the findings, it was confirmed that a high level of innovative work behavior is not present in companies in Bosnia and Herzegovina. The obtained results suggest the need for the development of more contemporary leadership styles and an organizational culture that encourages innovation.

TIG welding is used when a good weld appearance and a highest quality of the weld are required Nevertheless, the process has also some major disadvantages like relatively shallow penetration capability and low productivity. To increase the penetration and/or productivity, instead of traditional pure Ar, gas mixtures containing gases with high thermal conduction can be used. For austenitic stainless steels, as they are not prone to hydrogen cracking, also H2 is suitable. As H2 is active gas, the process is called Tungsten Active Gas (TAG). In this research, austenitic stainless steel sheet was welded with a competitive welding speed of 40 cm min−1, with pure argon and with Ar+7.5 H2 mixture. With pure Ar, a welding current of 220 A was far too small to reach full penetration in 3 mm sheet. With the 7.5 H2 active gas mixture, only 130 A was sufficient with unchanged welding speed.

A well-known characterization of Jordan vectors of a matrix polynomial $L(z)$ is generalized to a characterization of Jordan vectors of the operator-valued function $Q(z)$ at an eigenvalue $\alpha \in \mathbb{C}$. The results are then applied to solve a system of nonlinear ordinary differential equations.

This study examined the effects of structured dance programs on motor competence in preschool children. In a 12-week randomized controlled trial, 80 children were randomly assigned to two dance intervention groups (EG1: 2 × 35 min/week; EG2: 3 × 25 min/week) or a control group (CG) following the standard physical education curriculum. Fine motor skills, fine motor integration, and balance were assessed using the Bruininks–Oseretsky Test of Motor Proficiency, administered both before and after the intervention. Significant improvements were observed in EG1 for fine motor precision (p < 0.001), fine motor integration (p = 0.022), and static balance (standing on one leg on a balance beam with eyes open; p < 0.001). EG2 showed significant gains in dynamic balance (walking forward on a line; p < 0.001). Both dance programs enhanced preschoolers’ motor competence compared to the control, with higher session volume producing superior outcomes. These results support integrating structured dance sessions into preschool curricula to effectively enhance motor competence, offering a practical strategy to promote physical development in early childhood.

Learning control barrier functions (CBFs) offers a promising approach to enforcing safety but doing so with formal guarantees remains challenging. This is compounded by the complexity of the dynamics considered in this paper, given by systems in Brunovský canonical form, which naturally require high-order CBFs (HOCBFs). The signed distance function (SDF) naturally encodes HOCBF properties, but is often non-smooth, limiting its applicability in learning-based methods and safety-critical control. To address this, we propose a smoothing technique that ensures continuous differentiability with respect to the relative degree of the system dynamics while preserving key safety properties of the SDF. We then leverage Gaussian Process regression to learn an HOCBF candidate from noisy measurements, providing probabilistic safety guarantees through an inner approximation of the safe set. Additionally, we establish formal feasibility guarantees for the HOCBF-based controller and ensure the safety of the resulting closed-loop dynamics with high probability. Our approach enables online adaptation by efficiently updating the learned HOCBF with new data. We demonstrate our approach in simulation on a planar system and robotic manipulator with 2DOF.

In a rapidly growing market like the Albanian one, FinTech platforms are playing an increasingly important role in transforming how financial transactions are conducted. The research aims to perform a strategic assessment of ten FinTech platforms currently operating in Albania based on 12 strategic criteria. The assessment is based on multi-criteria decision-making methods (MCDM). For this, a model based on the fuzzy approach was developed, allowing for the management of uncertainty and subjectivity in evaluating the performance and suitability of the platforms. Specifically, the fuzzy LMAW method was used to weight the criteria, with Security being assigned the highest weight. The ranking of the platforms was conducted using the fuzzy CRADIS method, with EasyPay achieving the best results. Through this model, the research seeks to provide an objective ranking of the platforms based on each criterion’s relative contribution. The findings are expected to help developers, investors, and policymakers better understand the competitive positioning of current players. The results may also highlight areas for further improvement and growth in Albania's FinTech sector.

Multi-modal retrieval-augmented generation (MMRAG) promises grounded biomedical QA, but it is unclear when to (i) convert figures/tables into text versus (ii) use optical character recognition (OCR)-free visual retrieval that returns page images and leaves interpretation to the generator. We study this trade-off in glycobiology, a visually dense domain. We built a benchmark of 120 multiple-choice questions (MCQs) from 25 papers, stratified by retrieval difficulty (easy text, medium figures/tables, hard cross-evidence). We implemented four augmentations—None, Text RAG, Multi-modal conversion, and late-interaction visual retrieval (ColPali)—using Docling parsing and Qdrant indexing. We evaluated mid-size opensource and frontier proprietary models (e.g., Gemma-3-27BIT, GPT-4o family). Additional testing used the GPT-5 family and multiple visual retrievers (ColPali/ColQwen/ColFlor). Accuracy with Agresti–Coull 95% confidence intervals (CIs) was computed over 5 runs per configuration. With Gemma-3-27BIT, Text and Multi-modal augmentation outperformed OCR-free retrieval (0.722-0.740 vs. 0.510 average accuracy). With GPT-4o, Multi-modal achieved 0.808, with Text 0.782 and ColPali 0.745 close behind; within-model differences were small. In follow-on experiments with the GPT-5 family, the best results with ColPali and ColFlor improved by 2% to 0.828 in both cases. In general across the GPT-5 family, ColPali, ColQwen, and ColFlor were statistically indistinguishable; ColFlor matched ColPali while being far smaller. GPT-5-nano trailed larger GPT-5 variants by roughly 8-10%. Pipeline choice is capacity-dependent: converting visuals to text lowers the reader burden and is more reliable for mid-size models, whereas OCR-free visual retrieval becomes competitive under frontier models. Among retrievers, ColFlor offers parity with heavier options at a smaller footprint, making it an efficient default when strong generators are available.

This paper presents a comprehensive evaluation of multilingual cybersecurity text classification using conventional machine learning (ML) models, sentence-transformer embeddings, and open-source large language models (LLMs). We construct a manually labeled dataset of English and German news articles and benchmark models across zero-shot and fewshot settings while accounting for LLM knowledge cutoffs. Our results show that classic ML models, when combined with highquality embeddings, achieve performance equal to or better than state-of-the-art LLMs. For instance, an Multi-Layer Perceptron (MLP) classifier with multilingual-e5-large embeddings reaches an F1-score of 0.99 in the pre-cutoff setting, matching Qwen2.5-72B's few-shot performance ($F 1=0.99$) post-cutoff. Notably, this level of performance is achieved with over 99% lower computational requirements. Several embedding-based ML pipelines outperform all zero-shot LLMs, highlighting their costefficiency and robustness. These findings challenge the presumed superiority of LLMs and underline the importance of cutoffaware evaluations in practical applications.

The paper addresses the multifaceted challenges faced by post-mining regions, including environmental degradation, resource management, and socio-economic revitalization. Focusing on the city of Prijedor (Republic of Srpska, Bosnia and Herzegovina), the study proposes a holistic approach to land reclamation that integrates geoscientific data, hydrogeological analysis, and artificial intelligence (AI) techniques to identify optimal locations for spa and wellness development. The methodology combines fuzzy logic and evolutionary algorithms to process uncertain and multi-criteria datasets, such as aquifer characteristics, mineral water composition, geological formations, and infrastructure accessibility. The resulting model supports strategic planning by classifying zones based on their suitability for water resource exploitation and therapeutic use. Beyond ecological and economic considerations, the research highlights the importance of psychophysical health and complementary medicine. The envisioned transformation of former mining sites into “oases of wellness” includes facilities for yoga, meditation, reiki, and balneotherapy, promoting holistic regeneration of both land and community. This interdisciplinary framework exemplifies how AI-driven geoscientific strategies can foster sustainable development, turning degraded landscapes into centers of health, tourism, and environmental balance – honoring the past while shaping a resilient and innovative future.