Search Results (3,882)

The growing integration of chatbots and large language models (LLMs) into society raises important concerns about their potential to reproduce toxic human behaviors. As a result, it is essential to investigate these models to mitigate or eliminate such risks. This paper presents an experimental study evaluating the responses of the Llama and Alpaca models to scenarios involving verbal harassment. The methodology involved using harassment dialogues generated by an LLM as prompts to elicit responses from both models. The responses were then analyzed for levels of toxicity, sexually explicit content, and flirtatiousness. The results indicate that although both models reduce explicit offensive terms, they exhibit limitations in identifying and intercepting abusive behavior from users. Statistical analysis reveals that general-purpose instruction tuning in Alpaca does not provide a robust safety barrier compared to the Llama base model for most variables investigated in the experiment. However, a significant difference was observed concerning flirting, where Llama proved more prone to validation and encouragement than Alpaca. Furthermore, the study identifies critical vulnerabilities, such as a “self-deprecation” bias in Llama and “mirroring” behavior in Alpaca. We also report a complementary triangulation with GPT-family models as a secondary point of reference. This paper discusses and contains content that can be offensive or upsetting. Full article

Fire detection is a critical task for early warning systems, particularly in environments where visual sensing is unreliable. While most existing approaches rely on image-based or smoke-based detection, acoustic signals provide complementary information capable of capturing early combustion-related events. This study investigates deep learning models for sound-based fire detection, focusing on convolutional and Transformer-based architectures. VGG16 and VGG19 convolutional neural networks are adapted to process time-frequency audio representations for binary classification into Fire and No-Fire classes. An Audio Spectrogram Transformer (AST) is further employed to model long-range temporal dependencies in acoustic data. Finally, a hybrid VGG19-AST architecture is proposed, in which convolutional layers extract local spectral–temporal features, and Transformer-based self-attention performs global sequence modeling. The models are evaluated on a curated dataset containing fire sounds and diverse environmental background noises under multiple noise conditions. Experimental results demonstrate competitive performance across convolutional and Transformer-based models, while the proposed hybrid VGG19-AST architecture achieves the most consistent overall results. The findings suggest that integrating convolutional feature extraction with self-attention-based global modeling enhances robustness under complex acoustic variability. The proposed hybrid framework provides a scalable and cost-effective solution for sound-based fire detection, particularly in scenarios where visual monitoring may be obstructed or ineffective. Full article

This study presents the results of a Proof of Concept developed within the framework of the Amygdala Project, aimed at exploring the relationship between electronic music and emotional well-being among young people with and without a diagnosis of adjustment disorders (anxiety, depression, and distress). The fieldwork was conducted during the live concert of DJ Steve Aoki (Cosquín Rock 2024, Valladolid), combining psychophysiological measurements using Sociograph technology, self-reported questionnaires, and performative and contextual analyses. The results reveal significant differences between the two groups: participants with a diagnosis exhibited a more constant and profound emotional connection, interpreting the experience as a form of “emotional escape” and an opportunity for affective regulation; whereas those without a diagnosis experienced more fluctuating levels of attention and perceived the event primarily as entertainment. The triangulation of biometric, observational, and narrative data suggests that electronic music in collective contexts may operate as a tool for emotional containment and transformation, fostering group cohesion and reducing psychological distress. These findings open new avenues for interdisciplinary research into the biosocial effects of contemporary music and its potential in the design of cultural and educational strategies to promote psychological well-being among young people. Full article

In the study of eggplant (Solanum melongena L.), a cross between the green peel line 19143 and the white peel line 19147 produced E4957 F₁ hybrids with a purple–brown peel. Self-fertilization of the F₁ hybrids yielded E4957 F₂ offspring with a segregation ratio of 27:9:21:7 among individuals with purple–brown, purple–red, green, and white peel colors, respectively, which was consistent with a genetic model controlled by reciprocal recessive epistasis between D and P, and Gv1 likely acting as a modifying factor. The green peel line 19143 exhibited higher chlorophyll but lower anthocyanin levels than the white peel line 19147, which contained low levels of both pigments, while the E4957 F₁ hybrids had elevated levels of both pigments. Two epistatic genes, D and P, associated with anthocyanin synthesis, were mapped on chromosomes 10 and 8, respectively. The putative modifying locus Gf, involved in chlorophyll accumulation in the flesh, was mapped on chromosome 8, and the localization interval was close to the previously reported Gv1 locus associated with chlorophyll synthesis in the peel. DNA markers (InDel22522, InDel5531, InDel-APRR2) were developed to genotype 237 F₂ individuals and correlate genotypes with phenotypes. Sequence analysis revealed a 6 bp deletion in the SmMYB1 (D) gene and a large deletion in the SmAPRR2-Like (Gv1) gene in the white peel line 19147, as well as a T to A mutation in the SmANS (P) gene in the green line 19143. This study provided evidence for inheritance between loci involved in anthocyanin and chlorophyll pathways contributing to eggplant peel color variation and provides molecular markers that may facilitate the breeding of eggplant varieties with diverse peel colors. Full article

Copper-based hybrid metal matrix composites reinforced with graphene and zinc were developed to achieve a balanced combination of mechanical strength, corrosion resistance, wear performance, and electrical conductivity. In this study, Cu matrix composites containing a constant graphene content of 1 wt.% and varying Zn contents (0, 5, 10, and 15 wt.%) were fabricated through mechanical alloying followed by Spark Plasma Sintering (SPS). The effects of zinc content on microstructure, densification, hardness, corrosion behavior, tribological performance, and electrical conductivity were systematically investigated. Microstructural analyses revealed that the combined use of graphene and Zn significantly influenced grain refinement, interfacial stability, and densification behavior. The composite containing 10 wt.% Zn exhibited the highest relative density (~90.5%) and maximum hardness (62 HB), indicating an optimal reinforcement level. Corrosion tests conducted in 3.5 wt.% NaCl solution demonstrated that the 10 wt.% Zn composite showed the most noble corrosion potential and the lowest corrosion current density, which was attributed to reduced porosity and improved microstructural homogeneity. Tribological results confirmed that graphene contributed to a self-lubricating effect, while Zn enhanced load-bearing capacity, leading to improved wear resistance under increasing normal loads. Electrical conductivity measurements showed a gradual decrease with increasing Zn content, mainly due to solid-solution-induced electron scattering in the Cu matrix; however, the fixed graphene addition and effective SPS consolidation helped preserve conductive pathways, allowing all composites to retain acceptable conductivity levels. The results indicate that the hybrid Cu–graphene–Zn composites exhibit a balanced combination of mechanical, corrosion, tribological, and electrical properties, with 10 wt.% Zn emerging as the optimal composition. Full article

Polymer adhesive materials have been utilized across a wide range of applications, including adhesion to wood, metals, and biomaterial substrates. To meet increasing performance demands, the development of high-performance adhesive materials continues to be actively pursued by introducing advanced functions and capabilities into polymer networks. By incorporating dynamic covalent bonds into the polymer network, these materials gain self-healing and reprocessing abilities. While these materials exhibit high mechanical robustness and stability under service conditions, the bonding/rebonding reactions of dynamic covalent bonds allow the polymers to detach from target surfaces when needed. Additionally, noncovalent interactions within the network and between the polymer and the target surface significantly contribute to overall adhesive strength. Although dynamic covalent bonds and noncovalent interactions operate through different mechanisms, both contribute significantly to adhesive performance. This review manuscript presents studies on polymer networks containing dynamic covalent bonds and non-covalent interactions. Based on these studies, the respective contributions of each type of bond to the superior adhesive strength of the materials are discussed, and potential target substrates for adhesion, including wood, metal, and biomaterials, are proposed. Full article

Reverse osmosis concentrate (ROC) contains relatively high levels of refractory organic pollutants, posing significant challenges due to its difficult treatment and high environmental risks. Therefore, efficient and convenient removal strategies are essential. In this study, a self-developed iron-modified activated carbon fiber (Fe-ACF) was employed as an adsorbent to remove organic pollutants from ROC. Additionally, response surface methodology (RSM) was applied to model the adsorption process, identify and evaluate key influencing parameters, and optimize operational conditions. The adsorption mechanisms and regeneration stability of Fe-ACF were also investigated. Kinetic analysis revealed that the adsorption process is predominantly governed by chemisorption, with intraparticle diffusion identified as the primary rate-limiting step. Isothermal adsorption studies demonstrated that the Langmuir–Freundlich model best describes the adsorption behavior, yielding a theoretical maximum adsorption capacity of 12.21 ± 0.80 mg/g. Thermodynamic analysis confirmed that the adsorption process is spontaneous, endothermic, and driven by an increase in entropy. The RSM optimization identified pH as the dominant factor. The optimal adsorption conditions were a pH of 4.18, a temperature of 34.63 °C, a stirring speed of 547.91 rpm, and an adsorbent dosage of 1.55 g/L. The adsorption mechanism involves hydrogen bonding, π–π interactions, surface complexation, and electrostatic forces. Fe-ACF exhibits competitive regeneration stability and structural integrity. In summary, Fe-ACF demonstrates significant potential as a treatment material for ROC. Full article

The incorporation of reclaimed asphalt pavement (RAP) in asphalt mixtures improves sustainability but significantly reduces the intrinsic self-healing capacity due to binder aging. This study aimed to quantify whether epoxy-coated rejuvenator microcapsules could restore and enhance the self-healing performance of RAP-containing recycled asphalt mixtures. Four mixture types (AC-10C, AC-13C, AC-16C, and SMA-13C) containing 20% RAP were evaluated using a fracture–healing–refracture bending test (Repair index, R_C) and a splitting healing strength ratio (S_HSR) test to determine the effects of healing time, temperature, and microcapsule dosage. R_C increased rapidly during the first 8 h of healing and then approached stabilization, with the growth rate falling below 2%, indicating 8 h as the practical optimum healing duration. R_C increased from 0 °C to 45 °C due to enhanced binder mobility and diffusion, and slightly decreased at 60 °C because temperature-induced softening reduced peak bending strength. The highest self-healing capacity was obtained at a microcapsule dosage of 4% (by RAP mass). Under the optimum healing condition (8 h and 45 °C), R_C increased by 10.38%–13.50% and S_HSR increased by 14.35%–25.27% compared with mixtures without microcapsules. Among the mixtures, SMA-13C exhibited the highest self-healing capacity, followed by AC-13C, AC-10C, and AC-16C. The contribution of this study lies in quantifying the healing enhancement in RAP-containing mixtures, identifying practical optimum healing conditions based on a growth-rate criterion, and demonstrating consistent trends between two healing indices across different mixture structures. Full article

Traditional hardware-based approaches for depth-of-field extension (DOF-E), such as optimized lens design or focus-stacking via layer scanning, are often plagued by bulkiness and prohibitive costs. Meanwhile, conventional multi-focus image fusion algorithms demand precise spatial alignment, a challenge that becomes particularly acute in applications like microscopy. To address these limitations, this paper proposed a novel single-image DOF-E method termed SDENet. The method adopts an encoder –decoder architecture enhanced with multi-scale self-attention and depth enhancement modules, enabling the transformation of a single partially focused image into a fully focused output while effectively recovering regions outside the original depth of field (DOF). To support model training and performance evaluation, we introduce a dedicated dataset (MSED) containing 1772 pairs of single-focus and all-focus images covering diverse scenes. Experimental results on multiple datasets verify that SDENet significantly outperforms state-of-the-art deblurring methods, achieving a PSNR of 26.98 dB and SSIM of 0.846 on the DPDD dataset, which represents a substantial improvement in clarity and visual coherence compared to existing techniques. Furthermore, SDENet demonstrates competitive performance with multi-image fusion methods while requiring only a single input. Full article

Geopolymer materials obtained through the alkaline activation of fly ash represent a promising alternative for reducing the environmental impact of the construction sector, which is currently dominated by cement use. This study aimed to develop self-cleaning geopolymer composites by incorporating TiO₂ nanoparticles. Specimens containing 1%, 3%, and 4% TiO₂ were prepared using alkaline solutions based on Na₂SiO₃ and NaOH (6 M or 8 M), at mass ratios of 1:1 and 2:1. The results indicate that the three analyzed factors—the NaOH solution concentration, the activator ratio, and the nanoparticle dosage—significantly influence density, mechanical strength, and water absorption. Increasing the NaOH concentration to 8 M led to slight densification, improved flexural and compressive strength, and reduced water absorption. Modifying the Na₂SiO₃:NaOH ratio produced similar densification effects but resulted in reductions in mechanical strengths. The addition of 1–3% TiO₂ increased density and mechanical performance while reducing water absorption, whereas 4% TiO₂ content had the opposite effect. Self-cleaning capacity was confirmed by up to ~90% degradation of Rhodamine B after five UV–artificial rain–drying cycles, compared to only 27.3% degradation for the control samples. Full article

Aesthetic medicine has progressed from the early 2000s fascination with bio-stimulation to the current dominance of hyaluronic acid (HA) fillers, prized for immediate, predictable, and reversible volumizing effects. Recently, demand for more natural results, stronger emphasis on skin quality, and increased post-pandemic self-scrutiny have renewed interest in regenerative strategies, sometimes called the “second wave of bio-stimulation.” This trend highlights the need for clearer terminology and a cautious, evidence-based reading of proposed biological mechanisms. This narrative review proposes a framework in which bio-regeneration denotes a hypothesized, controlled induction of physiological processes, fibroblast activation, collagen and elastin synthesis, extracellular matrix remodeling, and immune modulation, potentially producing sustained improvements in dermal structure and function beyond simple filling. Among emerging technologies, polyethylene glycol diglycidyl ether (PEGDE) cross-linking is reported to create a stable, flexible HA scaffold with homogeneous tissue integration, favorable rheology, thermal stability, and a reduced inflammatory profile, supporting safer multimodal use with energy-based devices. The framework is illustrated with PEGDE-crosslinked HA combined with low-concentration calcium hydroxyapatite (CaHA), exemplified by a PEGDE-HA filler containing CaHA microspheres plus glycine and L-proline. These formulations aim to deliver immediate correction via HA and delayed stimulatory effects possibly driven by gradual CaHA exposure and macrophage-associated signaling. Available clinical, imaging, and histological observations, including prospective ultrasound and biopsy assessments, suggest progressive dermal thickening and predominant type I collagen expression, without pathological inflammation or granuloma formation. Although evidence remains preliminary and largely non-comparative, findings are compatible with controlled remodeling and resolving inflammation; however, the underlying mechanism and any ‘regenerative’ versus ‘reparative’ classification require controlled comparative studies. Full article

This work places the invariant $d{s}^2$ at the center of the gravitational interaction, interpreting it not as a purely geometric object but as the differential of proper time, endowed with direct physical meaning. Starting from the extension of Fermat’s principle to massive particles—namely, the requirement that freely falling bodies follow trajectories that extremize proper time, which for timelike motion corresponds to a local maximum—and invoking the universality of Galilean free fall, we derive the form of $d{s}^2$ in a static gravitational field. Lorentz invariance then provides the natural framework to extend this result to systems involving moving matter. The invariant derived through this procedure matches the weak-field limit of General Relativity formulated in the harmonic gauge. Within this linearized regime, we show that the structure of the theory already contains the seeds of its nonlinear completion: any dynamically consistent extension to strong gravitational fields necessarily involves the Ricci tensor. From this viewpoint, Einstein’s field equations appear not as a postulated geometric law but as the unique covariant closure required to ensure energy–momentum conservation and the self-consistency of the gravitational interaction. Full article

The foundry industry produces over 66 million tons of mixed casting waste sand, containing toxic and harmful substances such as phenols and aldehydes, every year, which has caused serious soil pollution, water source pollution, and large amounts of CO₂ emissions. Green resource recycling and utilization are urgently needed. The hot method circulating fluidized bed furnace is currently the mainstream technology for the regeneration of casting waste sand. However, traditional equipment has a series of key technical bottlenecks, such as VOC (volatile organic compound) emissions, low yield of fine sand, poor stability of phase change sand, and uneven fluidization, which directly limit the effectiveness, large-scale promotion, and application of waste sand regeneration. This study, based on a self-designed experimental prototype, constructed models with different hood densities and inlet air velocity parameters. A CFD-DEM coupled model, combined with two turbulence models, was used for numerical simulations and experimental validation, and the optimal combination of fluidization parameters was determined. The study confirmed that the k–ω SST model is more suitable for precise simulation of such gas–solid two-phase flows. The research revealed quantitative relationships between key parameters and sand particle fluidization states, addressing the core problem of uneven fluidization in conventional bubbling furnaces and providing important guidance for the optimized design of new thermal cycle bubbling furnaces. It has significant engineering value for promoting the efficient resource utilization of foundry waste sand and the green and sustainable development of the industry. Full article

In this work, compatibilized poly(lactic acid)/poly(butylene adipate-co-terephthalate) (PLA/PBAT) blends were developed and characterized, to be potentially utilized as biodegradable self-healing matrices for composite laminates. Blends containing 10, 20 and 30%wt of PBAT and 0.5 phr of an epoxy-based compatibilizer were prepared by melt compounding and hot pressing. Rheological measurements showed that moduli and complex viscosity generally increased with PBAT content, while maintaining viscosity levels suitable for conventional melt-processing operations. FT-IR and FESEM analyses confirmed the formation of an immiscible but well-compatibilized morphology, characterized by a homogeneous dispersion of PBAT domains within the PLA phase. Mechanical tests revealed a decrease in tensile modulus (up to 44%), strength (up to 45%) and fracture toughness (up to 40%) with a PBAT content up to 30%wt. Self-healing was evaluated by measuring the fracture toughness (K_IC) recovery after thermal treatment at 140 °C. After healing, the blend containing 20%wt of PBAT exhibited a self-healing efficiency of 64% under impact conditions, which was attributed to the smoother fracture surface generated at an elevated strain rate that facilitated a more effective flow of the molten PBAT phase across the crack interface during healing. The formulation containing 20%wt of PBAT featured the best balance between mechanical performance and self-healing efficiency. Full article

The COVID-19 pandemic posed unprecedented challenges, requiring compliance with public health measures. Notably, guilt is a powerful motivator for rule adherence; however, different types of guilt could have fueled the decision to stay home. This study investigated how guilt propensity influenced Italians’ self-reported motivations for adhering to containment rules. The propensity to different types of guilt, namely deontological and altruistic, was assessed in a total of 393 participants (261 females, 66.4%; 132 males, 33.6%; M age = 34.4, SD = 12.6) in May 2020, between the first and the second phases of Italian lockdown. The survey assessed four guilt dispositions—Moral Norm Violation (MNV), Moral Dirtiness (MODI), Harm-based guilt (HARM), and Empathy-based guilt (EMPATHY)—alongside fear of COVID-19, trust in authorities, and motivations for rule compliance (e.g., protecting one’s own and others’ well-being, respecting authorities, and avoiding sanctions). MNV emerged as a positive predictor of prosocial, authority-based and personal motivations, whereas MODI predicted lower prosocial motivation. HARM selectively predicted prosocial motivation and was negatively associated with authority-based motivations, while EMPATHY negatively predicted self-focused motivations. Moderation analyses revealed small but significant interaction effects, indicating that fear of COVID-19 slightly amplified the influence of EMPATHY and attenuated the effect of HARM, whereas trust in authorities strengthened the link between EMPATHY and prosocial compliance and reduced the association between MNV and prosocial motivations. These findings suggest that compliance during the pandemic was shaped by distinct emotional–moral pathways and that the motivational impact of guilt depends on perceived threat and institutional trust, highlighting the relevance of specific guilt profiles in promoting cooperative and health-protective behaviors. Full article