Abstract: Automated fake news detection has advanced substantially through transformer-based classification, yet two critical gaps persist in the literature: static models degrade as misinformation tactics evolve, and high-performing systems rarely reach end users in accessible forms. This paper addresses both gaps through a system that couples RoBERTa-based classification with a post-deployment continuous learning pipeline and a browser-native Chrome extension. We curate a corpus of 70,556 unique articles from three established benchmark datasets—ISOT, WELFake, and the COVID-19 Constraint dataset—after eliminating 42.9% of initially gathered samples as cross-dataset duplicates. A systematic comparison of XGBoost (95.88%), DistilBERT (97.74%), and RoBERTa-base (98.51%) establishes the production model, with selection driven primarily by false negative rate: RoBERTa achieves 1.09%, a 69% reduction over XGBoost and 28% over DistilBERT. A documented vulnerability of transformer classifiers is susceptibility to formally-worded misinformation that mimics journalistic style. We construct a dedicated adversarial training set of 70 examples spanning health misinformation, suppression narratives, and election fraud claims, and demonstrate that targeted fine-tuning raises adversarial detection accuracy from approximately 40% to 95.71% while maintaining 98.60% accuracy on standard benchmarks—achieved through experience replay that prevents catastrophic forgetting. For deployment, ONNX INT8 quantization reduces model size from 500MB to 125MB without accuracy loss, enabling inference on free CPU infrastructure. A GitHub Actions pipeline collects fresh labeled articles nightly, and a FastAPI service running on Hugging Face Spaces serves predictions with 150–200 ms latency. A Chrome extension providing paragraph-level hover detection, LIME-based word attribution, source credibility scoring, and multilingual support across 19 languages makes the system accessible to non-technical users. End-to-end evaluation across 50 curated articles yields 98% accuracy; research-backed adversarial testing across seven categories achieves 91.7%, with perfect detection on adversarial attacks, AI-generated misinformation, temporal domain shifts, and multilingual content.

Abstract: A metal recycling facility scrapyard fire that burned for four days continuously in February 2020 in rural Minnesota resulted in firefighters from around Minnesota to mobilize and aid. Combusted material included cars, refrigerators, metals, glass, foam, insulation. Urine, blood and serum specimens were collected one day later. Parameters collected included: CBC with differential, BMP, blood heavy metals, urine heavy metals, and serum heavy metals. This massive and prolonged industrial fire provided an opportunity for biomonitoring of hazardous, and unique, exposures acutely, in concordance with concerns raised by the employees at risk. Initial analysis of these results did not find evidence of acute concern regarding the biomonitoring results. However, some of these results may portend the potential for long-term consequences such as the development of occupational cancers, especially if there was recurrent exposure in prior or proceeding fires.

Abstract: Background/Objectives: Coexistence of systemic sclerosis and sarcoidosis is very ra-re. Both are systemic autoimmune diseases with lung involvement but with different pathogenesis. In contrast to findings of mid- to upper-lobe interstitial lung disease (ILD) or/with hilar lymphadenopathy in sarcoidosis, the most common lung manifes-tation of systemic sclerosis is lower-lobe ILD, which is typically characterized by a nonspecific interstitial pneumonia pattern. Distinction between lung involvement re-lated to each disease is crucial due to different therapeutic approach Methods We present herein a serie of three overlap cases: two with sarcoidosis onset before the diagnosis of systemic sclerosis and the other with systemic sclerosis onset before sarcoidos. Results: A review of cases of concomitant sarcoidosis and systemic sclerosis is dis-cussed, including the pathophysiology of each disease with shared pathways leading to the development of both conditions in one patient Conclusions: The systemic sclerosis-sarcoidosis overlap is a high-risk phenotype. Early recognition and a personalized, aggressive therapeutic approach are essential to alter the natural history of these two converging fibrotic and granulomatous processes.

Abstract: Background: Hallucination the generation of factually incorrect, internally in consistent, or ungrounded content remains a critical barrier to safe LLM deployment in high-stakes domains. Existing detection methods typically require external knowledge bases, model ne-tuning, or cloud API access, limiting applicability in local inference contexts. Methods: We evaluate the Kerimov–Alekberli (K–A) information-geometric framework as a real-time, inference-time hallucination detector across six open-source LLMs deployed locally on Apple M5 Silicon via Ollama v0.23.2 (Q4_K_M quantisation). The K–A framework monitors the KL divergence between consecutive output distributions relative to a Fisher Information Metric (FIM)-derived threshold (τ = 0.065), triggering First-Passage Time (FPT) alarms when generation departs from the stable Riemannian output manifold. We evaluate 120 responses (6 models × 20 questions) drawn from three established benchmarks: HaluEval (14 questions; categories: Fact, Confuse, Date, Num, Trap), FEVER (4 questions; adversarial fact verification), and SimpleQA (2 questions; precise factual recall). All questions are classified as difficulty level Hard, targeting known LLM failure modes including o-by-one numerical errors, geographical traps, and disputed-attribution confounds. Results: The K–A framework achieves a session hallucination detection rate of 90.9% (20/22 hallucinated responses correctly flagged) with zero false positives on correct responses (0/98). Model-level hallucination rates vary dramatically: deepseekr1:latest (Qwen3 CoT architecture, 5.2GB) exhibits a 95% hallucination rate (19/20 questions) with 100% K–A detection; gemma3:27b (Gemma3, 17.4GB) and gemma3:latest (4.3B, 3.3GB) achieve 0% hallucination. Two K–A false negatives involve con dent factual errors below the KL threshold. Average KL divergence for hallucinated responses (D̅_KL = 0.068 ± 0.004) is significantly higher than for correct responses (D̅_KL = 0.042 ± 0.016). Conclusions: K–A achieves competitive hallucination detection without external knowledge bases, ne-tuning, or cloud infrastructure, processing each response in real time with negligible overhead. The deepseek-r1 result reveals a fundamental tension between chain-of thought reasoning depth and factual precision on concise queries that warrants systematic investigation.

Abstract: (1) Background: Indonesia faces the dual challenge of a high tuberculosis (TB) burden and increasing drug resistance. Conventional molecular diagnostics frequently fail to detect isoniazid resistance and nontuberculous mycobacteria (NTM). This study evaluates a domestic multiplex Open PCR system in Medan, Indonesia. (2) Methods: From July to November 2025, 1,569 sputum specimens from suspected TB pa-tients were analysed using the Indigen MTB/NTM/DR-TB Real-time PCR Kit Gen 2. (3) Results: Mycobacterial DNA was detected in 421 specimens (26.8%). Among these, 396 (94.1%) were drug-susceptible TB, while 16 (3.8%) showed resistance, predominantly INH mono-resistance (n=14; 0.89% of total). Additionally, 9 cases (2.1%) involved NTM or TB-NTM co-infections. Tertiary hospitals showed significantly higher positivity rates (33.5%) than primary care (18.9%; p < 0.001). TB status was significantly associated with male (p = 0.0052) and older age (p = 0.006), whereas resistance profiles and NTM distribu-tion were consistent across all demographic groups (p > 0.80). (4) Conclusions: The Open PCR system effectively identified INH resistance and NTM cases overlooked by standard rifampicin-only assays. By bridging diagnostic gaps across a decentralized referral network, this facilitates rapid and targeted therapy. Integrating multiplex domestic innovations into national diagnostic algorithms is essential for achieving Indonesia’s TB elimination targets.

Abstract: Promising results in the regeneration of skin lesions have been demonstrated with the use of natural (organic) products, such as dermal dressings made of polysaccharides (chitosan complexed with xanthan), as they promote a hydrated and thermally insu-lating microenvironment, allowing gas exchange; and sources rich in growth factors such as autologous platelet-rich plasma (PRPa), which contains transforming growth factor beta (TGF-β), vascular endothelial (VEGF) and platelet-derived (PDGF), re-sponsible for stimulating the inflammatory cascade and healing. The aim of this study was to evaluate the action of a polymeric membrane (chitosan, xanthan and β-glycan) and PRPa on healing in vivo, used alone or in combination. For the tests, rabbits un-derwent a surgical procedure to induce the lesion and were distributed into a control group (GC), membrane group (GM), PRPa group (GPa) and membrane group associated with PRPa (GMPa), evaluated at moments M0, M7, M14, M21 and M28 (28 days). Wound color and exudation, presence of infection and inflammation, formation of granulation, scarring and necrotic tissue, morphological and morphometric analysis were evaluated. Statistical analyzes of the results were performed using the Software R® software, adopting a significance level of 5%. While statistical differences between treatments in healing time were not significant (p>0.05), all wounds achieved 100% retraction by M21. Notably, at M7, PRPa alone and in combination with the membrane contributed to higher wound retraction percentages (29.71% and 21.65%, respectively) compared to the control group (16.96%). These findings suggest that the complexed membrane, alone or combined with PRPa, fosters a humid environment, gas exchange, and antimicrobial activity crucial for healing, with PRPa further enhancing early wound retraction. It is concluded that the treatment of experimental surgical wounds with biodressings such as PRPa alone or associated with a complexed membrane of chitosan, xanthan, and β-glycan significantly contributes to wound retraction, in addition to offering a propitious and indispensable environment for the healing cascade, such as a humid environment, gas exchange, and antibacterial action. Future studies should consider a larger number of animals per group, histological evaluation for global tissue assess-ment, and collagen quantification.

Abstract: In this paper, which is entirely unconditional, we prove a sharpened almost-all theorem with fully explicit effective constants for the restricted weighted Goldbach sum$$R_{a,q}(N)\mspace{6mu}: = \mspace{6mu}\sum_{\substack{p_{1} + p_{2} = N \\ p_{1} \equiv a\ (mod\ q)}}^{}(\log p_{1})(logp_{2}),\quad q \geq 1,\quad gcd(a,q) = 1,$$whose expected main term is $M_{a,q}(N) = C_{2}S(N)N/\varphi(q)$, where $C_{2} = 0.6601618\ldots$ is the twin-prime constant and $S(N)$ is the binary singular series.Our results are organised around four pillars. (I) We give a complete character-pair decomposition of the second moment of the error $E(N): = R_{a,q}(N) - M_{a,q}(N)$, extracting the exact diagonal constant $G/(2\varphi(q))$, where $G = \prod_{p > 2}^{}(1 + (p - 1)^{- 2}) \in \lbrack 1.41320886,1.41320899\rbrack$ is the Gallagher--Goldston constant. (II) We establish a uniform minor-arc $L^{4}$ bound$$\int_{\mathfrak{m}}^{}|S(\alpha)|^{4}\, d\alpha\mspace{6mu} \leq \mspace{6mu}\kappa_{safe} \cdot 2^{A} \cdot \frac{X^{3}}{(logX)^{A}},\quad\quad\kappa_{safe} = 4.40,$$by combining the complete Vaughan identity with the Bombieri--Vinogradov theorem in integral form, giving an explicit derivation of $\kappa_{explicit} = C_{V}^{2}c_{L^{2}} = 4.004$ before applying a rigorous $10\%$ safety margin. (III) We derive the effective almost-all theorem$$\#\left\{ N \leq X\text{ even}:|R_{a,q}(N) - M_{a,q}(N)| > C(A,q)\, N(logN)^{- 3} \right\}\mspace{6mu} \ll_{A,q}\mspace{6mu} X(logX)^{- A},$$with the explicit constant $K: = 2C(1,4) \leq 38.02$, obtained from $C(1,4) \leq 19.01$ via a Stechkin-type optimisation. (IV) We prove a Pintz-type exceptional-set bound on $\{ N \leq X:R_{a,q}(N) = 0\}$.Every statement in the main body carries the tag \[PROVED\]. No Generalised Riemann Hypothesis, no zero-density hypothesis, no ternary sum $W_{a,q}(n)$, no spectral input, and no Chen-type sieve are used anyware.

Abstract: Inhomogeneous magnetic field distributions in high-frequency planar transformers frequently cause severe localized thermal hotspots and elevated leakage inductance. Traditional interleaved winding designs rely heavily on empirical trial-and-error, which becomes computationally prohibitive for multi-layer parallel structures due to the factorial "curse of dimensionality." To address this bottleneck, this paper proposes a universal, data-driven optimization methodology. First, a quantitative one-dimensional prefix-sum model is established to correlate winding arrangements with spatial magnetomotive force (MMF) distributions, effectively simplifying the electromagnetic evaluation. Subsequently, a customized Genetic Algorithm (GA) framework, featuring physical-constraint-preserving operators such as Order Crossover (OX), is introduced to efficiently navigate the high-dimensional discrete search space. Using an extreme 26-layer complex parallel winding configuration (Np:Ns = 9:2) as a primary case study, the proposed GA method effectively bypasses over 1.5 million permutations, converging to the global optimum within 100 generations. The optimized structure achieves profound peak-shaving, drastically reducing both the peak MMF and total uncoupled magnetic energy area. This methodology provides a systematic, computationally lightweight EDA solution that fundamentally replaces empirical trial-and-error in the design of high-frequency magnetic components.

Abstract: Nonlinear dynamical systems often exhibit complex collective behavior arising from the interaction of multiple elementary modes. In this work we investigate the aggregation of a countable family of dynamical modes generated by Lotka-Volterra systems and study the resulting structure in a Hilbert space of observables. The classical Lotka-Volterra equations form a fundamental class of nonlinear models describing interacting populations through coupled differential equations, while Koopman operator theory provides a framework in which nonlinear dynamics can be represented as a linear evolution acting on observable functions.We show that a countable aggregation of such dynamical modes admits a well-defined limit in a Hilbert space when the coefficients belong to l2. The resulting aggregated observable evolves according to the associated Koopman semigroup, yielding a linear representation of the underlying nonlinear dynamics in the observable space. We further prove that the geometry induced by this aggregated dynamics admits a canonical class of equivalent metrics generated by coercive operators, ensuring that the stability topology of the system is independent of the particular metric chosen within this class.Finally, we illustrate the theoretical framework by introducing a social risk functional defined as a quadratic observable associated with the induced metric. This example demonstrates how application-specific quantities can naturally arise from the geometric structure generated by aggregated nonlinear dynamics.

Abstract: Objectives: This prospective, open‑label, non‑randomized study evaluated a specific combination of cystine, B vitamins, and minerals (Pantovigar^® vegan) in 34 women (aged 30-75 years) with diffuse hair loss (Modified Savin Score 1–3) for improvements in hair condition and product acceptance and safety following up to 6 months of intake of one test product capsule three times daily. Methods: Hair condition was assessed using phototrichogram analysis, and participants completed a product acceptance questionnaire. Blood nutrient levels associated with hair loss were measured at baseline, and at 3 and 6 months. Results: From baseline to 6 months, a significant increase of 2.35% in the rate of anagen (p = 0.012) and a decrease of 2.35% in the rate of telogen (p = 0.012) hairs resulted in an increase of 2.37% in the ratio of anagen/telogen hair rate (p = 0.013), reflecting a proportionate increase in the number of terminal hairs in the anagen phase. Most participants agreed with positive questionnaire hair condition statements. Increased levels of B vitamins, ferritin, and hematocrit did not exceed established tolerable upper limits, and no serious undesired effects were reported. Conclusion: The study results suggested that the specific nutrient combination of cystine, B vitamins, and minerals in the test product is beneficial to hair condition, is well‑accepted, and safe for use in the management of diffuse hair loss in women. Clinical Trial Registration: [https://clinicaltrials.gov/study/NCT05943860]; identifier [NCT05943860]; date of registration: June 21, 2023; retrospectively registered.

Abstract: The physiological acidity of the skin microenvironment constitutes a paramount component of passive innate defense against pathogenic colonization and subsequent dysbiosis. The objective of this review is to critically evaluate the sequelae of chronic epidermal barrier exposure to chemical exfoliants and to elucidate the impact of hydroxy acids on microbiota stability. An analysis of recent literature delineates fundamental discrepancies in their antimicrobial mechanisms, inflammatory modulation, and induction of stratum corneum desquamation. Conventional exfoliants, notably glycolic and salicylic acids, exhibit the capacity for passive diffusion into bacterial cells, where they undergo dissociation, precipitating destructive cytoplasmic acidification; furthermore, salicylic acid specifically downregulates crucial transcription factors. While these mechanisms facilitate the efficacious eradication of pathogenic organisms, they concurrently jeopardize commensal populations, which may paradoxically pave the way for secondary dysbiosis and opportunistic colonization. Next-generation acids, such as lactobionic acid, emerge as a highly promising, albeit insufficiently investigated, alternative. They exhibit the capacity to compromise bacterial cell walls and intercalate with microbial DNA. Future paradigms regarding the utilization of hydroxy acids in clinical dermatology must extend beyond traditional macroscopic endpoints to encompass the preservation of microbiome diversity and stability, thereby precluding chronic impairment of the associated protective barrier.

Abstract: This paper examines the transmission of monetary policy through bank credit by distinguishing between credit to non-financial corporations and household credit in an emerging economy. Using an ARDL–ECM framework with structural break analysis, it investigates the relationships between credit, the policy rate, and inflation over time. The results suggest that the conventional interest rate channel is limited, as the policy rate does not exhibit a statistically robust association with credit in either the short or long run. Inflation shows a differentiated and context-dependent association, remaining weak for corporate credit but more consistently related to household credit, which may reflect the role of nominal conditions in shaping borrowing behavior. Credit dynamics also differ across segments. Household credit appears more persistent and adjusts gradually, whereas corporate credit is less inertial but more sensitive to instability, pointing to heterogeneous adjustment processes. In addition, the estimated relationships are not stable over time, as their sensitivity to macroeconomic variables varies across periods. Overall, the findings indicate that monetary transmission is heterogeneous and time-varying rather than uniform. These results should be interpreted as conditional relationships within the empirical framework and provide a disaggregated perspective on how macroeconomic conditions are associated with credit dynamics in bank-based emerging economies.

Abstract: Background/Objectives: Eating disorders (EDs) are among the most severe psychiatric conditions affecting young people, with increasing prevalence in the post-pandemic period. This study assessed the prevalence of ED risk and dysfunctional eating behaviors among Italian university students, a population poorly characterized with respect to ED risk, and examined associations with key socio-demographic and anthropometric variables. Methods: A cross-sectional online screening study was conducted between August 2023 and February 2026 with 401 Italian university students (women: n = 306; men: n = 95). Participants completed the validated Italian versions of the Eating Attitudes Test-26 (EAT-26) and the Eating Disorder Examination Questionnaire 6.0 (EDE-Q 6.0), alongside self-reported anthropometric data. Multiple linear regression analyses were performed to identify predictors of ED risk scores. Results: 37.9% of participants had an abnormal BMI (19.7% underweight; 18.2% overweight or obese). EAT-26 scores exceeded the clinical cut-off in 28.4% of participants (women: 35.6%; men: 5.3%). EDE-Q 6.0 global scores exceeded the clinical cut-off in 21.0% (women: 25.8%; men: 5.3%). Only 45.4% showed no anthropometric or psychometric risk indicators. Gender was the strongest predictor of both EAT-26 and EDE-Q 6.0 scores. BMI was negatively associated with EAT-26 scores in the total sample and in women, while a positive association between BMI and EDE-Q 6.0 scores was observed in men. Conclusions: A substantial proportion of Italian university students, particularly women, presented clinically significant ED risk. The combined use of anthropometric and psychometric screening tools provides a more comprehensive risk assessment than either measure alone, highlighting the need for multidimensional screening programs.

Abstract: Palmatine chloride (berbericinine, C₂₁H₂₂ClNO₄) is a protoberberine alkaloid found in several plants including Rhizoma Coptidis, Cortex Phellodendri, Rhizoma Corydalis, Guduchi (Tinospora cordifolia), and Tinospora sagittata roots. Palmatine chloride (PA) is known as an inhibitor of dopamine generation. However, its effect on endoplasmic reticulum (ER) stress-related macrophage activation caused by endotoxin (lipopolysaccharide) is not well known yet. In this study, effects of PA on pyroptotic responses of mouse macrophages (RAW 264.7) activated by endotoxin were investigated using Griess reagent assay for nitric oxide (NO) production, fluo-4 assay for cytosolic calcium release, dihydrorhodamine 123 assay for hydrogen peroxide production, multiple cytokine assay for cytokines production, real-time PCR for inflammatory genes transcriptions, and flow cytometry assay for p38 MAPK activation. Results revealed that PA significantly reduced excessive production levels of NO, hydrogen peroxide, pro-inflammatory cytokines (such as interleukin (IL)-6, CCL3 (MIP-1α), and CSF2 (GM-CSF)), and cytosolic calcium release in endotoxin-stimulated RAW 264.7, but significantly increased the production of anti-inflammatory cytokine IL-10. PA inhibited endotoxin-induced transcripts of Chop, Stat1, Fas, and c-Fos in activated RAW 264.7. It also decreased p38 MAPK phosphorylation and level of Fas in RAW 264.7 stimulated by endotoxin. To further interpret these findings, a network pharmacology-informed analysis based on large-scale literature mining was performed, supporting the multi-target regulatory role of PA in ER stress-related pathways. Briefly, PA exerts anti-inflammatory effects on endotoxin-stimulated RAW 264.7 via calcium-CHOP pathway, consequently reducing endotoxin-induced production of pro-inflammatory mediators (NO, cytokines, etc.) and relieving ER stress-related pyroptotic cascade.

Abstract: Early 2026 has witnessed significant volatility in the oil market, and an energy crisis is expected in the coming months. Large-scale LLM inference continues to consume substantial power in data centers and improving inference efficiency is therefore increasingly important for energy sustainability of AI economy. Semi-structured N:M sparsity, most notably 2:4 (50%) and potentially 1:4 (75%), offers a hardware-friendly path to lower compute and energy, and has been supported in modern GPU designs. Yet existing training methods for 2:4 sparsity (e.g., STE-based approaches) often incur large accuracy drops relative to dense baselines, and practical support for 1:4 remains limited in current software stacks. As a result, attention has shifted toward quantization and mixture-of-experts, leaving high-sparsity N:M pre-training underexplored. Here we introduce a paradigm shift: we treat neural networks as complex systems whose sparse connectivity can be trained using network-science principles formalized by Cannistraci–Hebb sparse-to-sparse training (CHT), coupled with a tailored optimizer. We propose CHTsNM, a sparse-to-sparse training framework centered on Topology-Aware Newton–Schulz (TANS) optimization. TANS makes Newton–Schulz-style matrix updates compatible with dynamically changing semi-structured sparse topologies via active-mask projection, active-support RMS matching, and refresh-aware ramping after topology updates. CHTsNM further incorporates two lightweight mechanisms: Contextually Modulated LoRA (CoMoLoRA) for input-adaptive low-rank residual compensation, and Motif Pattern Revisitation (MPR) to improve exploration of legal row-wise N:M patterns. Across 4 LLaMA pre-training benchmarks, CHTsNM with 2:4 sparsity achieves performance close to dense baselines on most tasks and yields sparse-over-dense gains on 8 tasks. 1:4 sparsity approaches dense performance, though does not yet consistently surpass it. For hardware evaluation, we report measured speedups for native 2:4 execution on current NVIDIA GPUs, and provide a clearly labeled CSR sparse-GEMM surrogate analysis to estimate the acceleration potential of 1:4. Overall, although not implement on hardware yet, our results identify 1:4 sparse pre-training as a promising direction and establish TANS sparse-to-sparse optimization as a practical step toward future high-sparsity N:M accelerators.

Abstract: Forest edge effects are commonly interpreted as radial gradients from the edge toward the interior, but this assumption may oversimplify the spatial organization of heterogeneous tropical forest fragments. Here, we integrated field-based phytosociological data with Sentinel-2 spectral indices to evaluate whether edge-effect interpretation depends on analytical scale in Semideciduous Seasonal Forest fragments embedded in the Brazilian Cerrado. Five fragments were analyzed using transect-based plots and continuous pixel-level modeling. Basal area showed a strong positive correlation with NDVI (r = 0.95), supporting its use as the main spectral proxy for vegetation structure. Plot-level segmented regression detected edge-to-interior transitions, with breakpoints ranging from approximately 13 to 39 m. However, pixel-level modeling revealed scale-dependent responses, including shallow gradients in IF and AC, an intermediate transition in CN, a deeper gradient in IP, and high internal heterogeneity without a single dominant radial transition in Panga. The first two PCA axes explained approximately 81% of the total variance, reinforcing the structural–spectral correspondence. These findings show that edge effects are detectable but not adequately represented by fixed radial zones alone. Pixel-level Sentinel-2 modeling improves the spatial interpretation of fragmented tropical forests.

Abstract: Background: Epiretinal membrane (ERM) is a fibrocellular preretinal proliferation composed of fibroblasts, glial cells, and hyalocytes overlying the internal limiting membrane. Pars plana vitrectomy (PPV) with membrane peeling is the standard surgical treatment, but postoperative cystoid macular edema (CME) can limit visual recovery. Inflammation plays a key role in CME pathogenesis, and corticosteroids may help reduce the inflammatory response associated with both ERM-related traction and surgical trauma. Therefore sustained-release dexamethasone (DEX) intravitreal implants have been investigated as an adjunct to surgery. Methods: A narrative review of the literature was conducted using PubMed and Embase databases up to January 2025. Studies evaluating DEX implants in conjunction with PPV and ERM peeling were included. Both prospective and retrospective clinical studies were considered, in accordance with SANRA recommendations. Results: DEX implantation appears to promote faster resolution of postoperative inflammation and CME, with consistent improvements in anatomical parameters such as central macular thickness. However, the impact on visual acuity remains variable, with some studies reporting earlier functional recovery and others showing no significant long-term benefit. Reported adverse events mainly include intraocular pressure elevation manageable with medical therapy and cataract progression. Conclusions: DEX implant appears to be a safe and potentially effective adjunct to PPV with membrane peeling for ERM, particularly in high-risk eyes, although further large, prospective randomized studies are needed to better define its role.

Abstract: Accurate rainfall data are essential for hydrological forecasting and climate modeling. However, many developing regions, including Tunisia, struggle with significant data gaps in rainfall measurements, particularly from gauge stations. These missing data impair climate model validation and reduce forecasting accuracy across both spatial and temporal dimensions. To overcome these limitations, we conduct a comprehensive evaluation of novel deep learning (DL) architectures designed for imputing missing rainfall gauge data and generating monthly rainfall forecasts. Our framework systematically compares multiple DL approaches: Long Short-Term Memory (LSTM), a hybrid Bidirectional LSTM with a Transformer attention mechanism (BiLSTM-Transformer), and a pure Transformer model. Subsequently, we employ Principal Component Analysis (PCA), K-Means clustering, and quantile techniques to further refine DL model outputs. The processed data are then analyzed using Light Gradient Boosting Machine (LightGBM) to produce final results. Our rigorous evaluation across 47 Tunisian gauges covering 1983–2012 (70% training, 30% testing) demonstrates that the BiLSTM-Transformer hybrid delivers superior performance, achieving an 18.4% reduction in root mean squared errors (RMSE) compared to conventional interpolation methods (14.2 mm versus 17.4 mm monthly error) and improving R2 values by 0.15–0.23 across all test stations. The model shows particular strength in capturing Mediterranean rainfall patterns, correctly predicting 83% of extreme rainfall events (greater than 95th percentile). Furthermore, spatial graph networks boost performance at data-sparse stations by 12.7% through explicit modeling of topographic influences.

Abstract: For a Hutchinson iterated function system (IFS), a Banach contraction on a complete metric space, or a finite-metric dynamical system, a natural question is: at which resolution σ does the contraction's geometric structure (fractal attractor, basin of attraction, periodic part) become optimally visible? We answer this by introducing a scale-selection principle: define the observation scale σ_c := argmax_σ χ(σ), where χ(σ) = |dO(σ)/d log σ| is the susceptibility of a resolution-dependent observable, and prove that σ_c exists under explicit boundary-regularity hypotheses.The framework's main quantitative results are three theorems specialised to Banach contractions and IFS: (i) a geometric scaling identity σ_c = qL for affine Banach contractions with operator norm q and basin scale L, applying directly to Hutchinson IFS with σ_c ∼ q · diam(K⋆); (ii) a discrete Banach theorem on finite metric structures under uniform Lipschitz Lip_d(f) = q < 1, giving an exact collapse-time N⋆ = ⌈log(∆/d_min) / log(1/q)⌉; (iii) a spectral concentration theorem placing σ_c at the inverse log-spectral-gap of the transfer operator at fixed positive noise. A stability lemma for canonical normalisation under smooth windowing and a parametric Banach correspondence observation complete the technical core.The framework is stated explicitly in the non-expansive Lipschitz regime Lip_d(f) ≤ 1 on the metric side, and at fixed positive noise ε ∈ (0, 1) on the spectral side. A four-type classification of operations by injectivity structure organises the broader landscape; cross-domain empirical evidence anchored on a peer-reviewed NISQ-hardware measurement of σ_c is summarised. The middle-thirds Cantor set IFS appears as the principal worked example.

Abstract: The study applies a multiphase, multimethod research approach based on the participatory methodology. It integrates perspectives of professionals in the travel industry and academic experts with the aim to develop an integrated conceptual model of the AI and IoT influence on work, skills development, and job attractiveness in the industry. The research provides a comprehensive understanding of the ways in which digital technologies indirectly shape employment through changes in work processes and development of transferable digital and socio-emotional skills. The paper aimed to donate to redefining the perception of work in tourism and hospitality, emphasizing the sector not only as a career choice but also as a platform for the acquisition of skills relevant in other industries as well. The outcomes revealed that the employees’ aspirations to enter or stay in the industry are not directly influenced by AI and IoT technologies; rather, their effects are mediated through changes in work processes and, more importantly, through the development of skills. The study contributes theoretically by evolving and analytically confirming an incorporated theoretical model that connects technology implementation, work transformation, skills development, and employment outcomes. Practically, the results underscore the importance of human-centered implementation strategies, emphasizing training, communication, and employee inclusion to maximize the benefits of digital technologies.