The West University of Timișoara (Romanian: Universitatea de Vest din Timișoara; abbreviated UVT) is a public higher education institution located in Timișoara. Classified by the Ministry of National Education as a university of education and scientific research, UVT is one of the nine members of the Universitaria Consortium (the group of Romanian elite universities). Also, the West University is a component institution of the National Research–Development–Innovation System in its capacity as an accredited higher education institution.The university was founded by the Decree-Law no. 660 issued on 30 December 1944, which stipulates that a university must be created in western Romania. Its first faculties either were dissolved or became independent institutions. These independent higher education institutions became a sole university at the end of September 1962. In 1968 the institution became an independent university. A difficult period followed, especially for the humanities and the exact sciences. Fields of study such as music, fine arts, history, geography, natural sciences and chemistry have all but disappeared, while philology has greatly restricted its activity. As a result, many of the faculties were transferred to other institutions. 1989, the year of the Romanian Revolution, was a turning point in the evolution of the university. It led to the present-day conditions. An essential mentality change took place in the perception of the academic institution, in accordance with the democratic institutions in Western Europe. A substantial institutional reform was carried out, beginning with the redefining of the university's mission by establishing its objectives.The West University of Timișsoara comprises 11 faculties with the respective departments, as well as a teacher training department. The faculties that operate within UVT offer nationally accredited study programs at bachelor's, master's and doctoral level in the following fields: Arts and Design; Chemistry, Biology, Geography; Economics and Business Administration; Law; Letters, History and Theology; Mathematics and Computer Science; Music and Theater; Physical Education and Sports; Physics; Political Sciences, Philosophy and Communication Sciences; Sociology and Psychology.
Multi-horizon solar irradiance nowcasting is highly sensitive to rapidly varying atmospheric conditions. Traditional statistical models such as persistence and autoregressive integrated moving average are computationally efficient but often degrade under strong atmospheric variability, whereas physical models based on numerical weather prediction remain constrained by spatial resolution, initialization delay, and computational cost. Machine learning methods such as extreme gradient boosting (XGBoost) improve nonlinear regression performance, but when used alone they do not explicitly capture temporal dependencies. Deep learning models, including convolutional neural network (CNN) and temporal convolutional network (TCN), can learn temporal patterns, yet maintaining stable accuracy across multiple forecast horizons remains challenging. To address this issue, this study proposes a hybrid CNN–TCN–XGBoost framework in which CNN and TCN extract multi-scale temporal features and XGBoost performs horizon-specific regression. The model is evaluated over forecast horizons from 1 to 30 min using independent 2018 and 2021 datasets covering clear-sky, mixed, overcast, and highly variable atmospheric regimes. The results show notable improvements over the 2-State benchmark, including normalized root mean square error (nRMSE) reductions of approximately 8% at 2 min and 22% at 10 min in the aggregated 2018 evaluation, while preserving smooth error growth with increasing horizon. Day-level validation across six representative atmospheric regimes provides further evidence of regime-dependent forecasting performance: on a highly variable day, error reductions reach about 38% at 10 min and 29% at 30 min, while on a stable clear-sky day the reduction reaches nearly 87% at 30 min. For the clear-sky subset, nRMSE is typically below 0.016 and remains below 0.032 even on the most challenging day. Therefore, the proposed CNN–TCN–XGBoost framework is best interpreted as a model that provides strong short- and intermediate-horizon performance and best or near-best accuracy across most evaluated regimes, while long-horizon performance remains dependent on atmospheric regime across diverse atmospheric conditions. These findings make it a promising solution for reliable real-time solar irradiance forecasting and photovoltaic operation.
PurposeThis study examines how Human-AI Collaboration (HAIC) and Cognitive Flexibility (CF) shape decision-making in the emerging context of Industry 6.0. It conceptualizes hybrid intelligence as a governance issue in which the level of AI integration and the adaptive judgment of managers develop separately and also influence one another in project and innovation settings.Design/methodology/approachUsing the scenario method, the study identifies HAIC and CF as two main uncertainties in AI-mediated decision-making. A two-dimensional framework is used to develop four configurations: Agile Synergy, Intuitive Mastery, Tech-Driven Adaptability, and AI-Controlled Rigidity. Each scenario captures a different combination of AI integration and cognitive adaptability.FindingsThe analysis suggests that decision quality in AI-enabled projects depends on the fit between computational capability and human interpretive capacity. High levels of both HAIC and CF support adaptive governance and more responsible decision-making. Imbalances create different risks, including procedural dependence and weaker strategic foresight, as interpretive authority gradually shifts toward system outputs. The findings also show that technological sophistication alone does not ensure resilience or ethical alignment.Practical implicationsThe framework helps project leaders and innovation managers assess how deeply AI is embedded in decision processes and whether decision-makers still have the capacity to interpret and challenge algorithmic outputs. It is also relevant for digital transformation teams involved in the development and use of AI-enabled decision systems in areas such as project selection, innovation portfolio review, risk assessment, resource allocation, and stage-gate processes, where explainability and reflective human judgment remain important.Social implicationsBy placing AI integration within a humanistic governance perspective, the study highlights the importance of preserving human interpretive agency and maintaining accountability to stakeholders in projects shaped by advanced algorithmic systems.Originality/valueThe study contributes to research on project and innovation management by showing that HAIC and CF capture two aspects of hybrid intelligence: one structural and one cognitive. It treats HAIC as a structural dimension of authority distribution and CF as a safeguard for judgment in hybrid systems. It also presents hybrid intelligence as a relational configuration and offers a governance perspective on digital transformation in project contexts.
Increasing self-esteem through an online intervention may be essential for students' psychological well-being and academic performance. Based on Melanie Fennell's cognitive model (1997, 2016), this study investigates the feasibility of an online intervention program aimed at improving students' self-esteem. After completing online screening, 33 students (M-age = 22 years, SDage = 6.75; 70% females) with low or medium levels of self-esteem participated in this open-label uncontrolled pilot trial. The primary outcome measure was the Rosenberg Self-Esteem Scale (RSE). As part of the intervention program, we utilized psychoeducation and cognitive-behavioral techniques. Furthermore, due to the pandemic context in which the intervention's feasibility was tested, we delivered the intervention online, in working groups. Following this five-week program, participants exhibited higher levels of explicit self-esteem (d = .74) and lower levels of depression and anxiety as secondary outcomes (d = .46 for depression and d = .48 for anxiety). Despite inherent limitations, these results provide initial evidence supporting the feasibility of an online working group intervention aimed at improving students' self-esteem.
A cross-sectional survey was conducted with 202 Romanian adults (aged 18-61), who completed validated measures assessing personality traits (Big Five agency, beliefs, conscientiousness, dynamism, and morality), cognitive-emotional coping strategies (cognitive emotion regulation questionnaire), and the use of AI, IoT, and blockchain technologies (hours per day). Data were analyzed using Jamovi, applying descriptive statistics, correlations, multiple regression with Bonferroni corrections, and mediation/ moderation analyses with bootstrap resampling. The analyses indicated no evidence of common method bias. Among the three tested models, only AI use was significantly predicted by personality factors, with extraversion exerting a positive effect and maturity a negative effect. Age moderated the extraversion-AI relationship, suggesting stronger effects among younger participants. Mediation analyses showed that adaptive coping strategies did not play a significant mediating role. Personality factors, particularly extraversion and maturity, play a central role in the adoption of AI, while coping strategies showed limited explanatory power. The moderating effect of age suggests that younger individuals may benefit more from extraversion in engaging with digital technologies. These findings underscore the importance of considering
Abstract Advances in taxonomy often uncover cryptic species nested within already protected taxa, raising concerns about the continuity of legal safeguards. While newly described species are not necessarily disconnected from existing protections, often being treated administratively as part of the original species complex, formal legislative updates remain essential to avoid confusion and ensure clarity. This tension, sometimes referred to as taxonomic anxiety, challenges researchers to balance scientific progress with conservation certainty. Using the case of the Idle Crayfish (Austropotamobius bihariensis), an endemic species recently described in Romania, we demonstrate how robust scientific evidence, coordinated stakeholder engagement, and strategic communication enabled full legal recognition. We propose an actionable roadmap for integrating taxonomic advances into conservation policy, offering a replicable model for ensuring legal protection of newly recognized species worldwide. While our case study focuses on the EU Habitats Directive framework, the challenges and solutions we present are relevant to conservation policy systems globally.
