Abstract: This paper proposes an anomaly detection algorithm based on large language models guided by knowledge graphs. The method combines the deep semantic understanding capability of large language models with the structural modeling advantages of knowledge graphs, forming a semantic-structural co-driven fraud detection framework. Specifically, the model first encodes transaction texts using a pretrained language model to generate semantic representations. Meanwhile, a financial knowledge graph is constructed based on entities such as transaction accounts, devices, and IP addresses, from which structural embeddings are obtained via graph neural networks. Next, a gated fusion mechanism is introduced to adaptively integrate the semantic and structural vectors. The fused representation is then fed into a classifier to produce fraud prediction results. In multiple comparative experiments, the proposed model outperforms single-modality approaches across metrics including accuracy, precision, recall, and F1-score, validating its effectiveness in complex financial scenarios. In addition, this study conducts multi-task adaptation experiments and anomaly ratio variation tests to evaluate the model's stability and robustness. The results show that the proposed method maintains strong performance across different environments, demonstrating its reliability and practical value.

Abstract: Traditional drying methods often result in significant loss of flavor, color, and bioactive components in food products. One effective method of improving the nutritional value of foods and shortening the drying process is to apply heat via continuous microwave drying. Understanding continuous microwave drying of crab apples is key to boosting efficiency and energy savings. In this study, crab apple slices were dried under continuous microwave drying and thoroughly investigated in terms of dehydration characteristics and quality attributes. The effects of drying conditions, including slice thickness (1–5 mm), microwave power (11400–19000 W), air velocity (0–2 m/s), and drying time (12–16 min) on the changes in temperature, moisture content, color, texture, total phenolic content, and microstructure were analyzed. Results showed that increased microwave power and decreased slice thickness significantly reduced moisture content; increasing air velocity reduced the temperature increase, whereas prolonging drying time decreased moisture content and increased drying temperature. As the slice thickness increased, the color values showed a decreasing trend while texture and total phenolic content increased. The application of higher microwave power resulted in changes in the overall color and texture, which decreased while the total phenolic content increased, additionally damaging the surface structure. Higher air velocity increased the color and texture while reducing the total phenolic content. Moreover, heightened air velocity may lead to more apparent structural changes. As the drying time increased, the color and hardness values decreased while the total phenolic content increased; prolonged drying times can significantly change the microstructure. To ensure the quality of the dried product, it is recommended that high microwave power, excessive air velocity, and extended drying times be avoided. Therefore, the appropriate conditions for continuous microwave drying of crab apple slices are a slice thickness of 1.5 mm, microwave power of 15200 W, air velocity of 0.5 m/s, and drying time of 13 min. This study may offer guidance for understanding and improving the continuous microwave drying of dried products, such as crab apples.

Abstract: Researchers show a lot of interest in the probe of graph invariants with respect to graph products as it poses challenging questions and provides interesting insights. The graph invariants considered here are domination, total domination and captive domination numbers. Let G = (V, E) be a simple undirected graph. A set D⊆V(G) is called a dominating set if any v ∈ V(G) – D is adjacent to at least one element in D and the size of a minimum dominating set of G is called the domination number of G denoted by γ(G).D is called a total dominating set if any v ∈V(G) is adjacent to at least one element in D and the size of a minimum total dominating set of G is called the total domination number of G denoted by γt(G).The graph product taken up for investigation is the strong product of two path graphs. A motivation for this stems from the conjecture: Let Γ = CNN [n, m] be the (n, m)- dimensional CNN such that m, n ≥ 2. Then γΓ = n3m3raised in [1]. Here the CNN denotes cellular neural networks. In the language of graph theory, the Γ is isomorphic to the strong product of two path graphs denoted byPn⊠ Pm. In this paper, while disproving this conjecture for certain cases we also found the exact values of γΓ for all n and m by carefully analyzing the underlying structures for connection pattern. While doing so, we also end up with the exact values of γtPn⊠Pm and a few other results.

Abstract: Classification of BIM objects is critical for enhancing information interoperability and standardization within construction projects; however, research on automated BIM object classification based on standardized classification systems remains limited. Therefore, this study aims to develop an automated classification method based on the Uniclass system, utilizing low-LOD IFC data combined with machine learning algorithms to achieve standardized classification of BIM objects. The proposed method first assigns Uniclass codes to BIM objects, then extracts IFC data and coding information as feature variables and classification labels, and finally applies a Random Forest model to perform automatic classification. Experimental results demonstrate that the proposed method achieves classification accuracies of 1.00 and 0.99 for BIM objects at the Elements/Functions and Systems coding tasks. This model enables the automatic classification of BIM objects relying solely on low-LOD input, thereby reducing dependence on high-detail modeling and improving applicability during the early design phase.

Abstract: The increasing frequency of cyber threats and the enforcement of strict privacy regulations have exposed critical limitations of traditional centralized machine learning models, especially in distributed environments such as the Internet of Things (IoT). This study presents a federated learning (FL) framework tailored for intrusion detection and malware classification that enables decentralized model training while preserving data locality and minimizing communication overhead. The proposed architecture incorporates lightweight privacy-preserving techniques-including gradient clipping, differential privacy, and encrypted model aggregation-to ensure secure and efficient collaboration across heterogeneous clients. Experimental results on benchmark datasets, such as CICIDS2017 [1] and TON_IoT [2], show that the framework achieves detection accuracies above 90%, while maintaining privacy loss below 5% and improving communication efficiency by more than 25%. These results confirm the viability of federated learning as a scalable and privacy-compliant approach for next-generation cybersecurity systems in highly distributed infrastructures.

Abstract: The black hole information paradox persists because no existing framework shows how the encoded information becomes operationally accessible to a physical observer. Starting from Tomita–Takesaki modular flow, we derive an observer-dependent retrieval law: dS_retrieved/dτ = γ(τ) [S_max − S_retrieved(τ)] tanh(τ / τ_char) This converts global entropy conservation into a Lorentzian-causal, time-resolved recovery process. The law predicts class-specific trajectories and an acceleration-dependent g^(2)(t1, t2) interference envelope detectable in current Bose–Einstein condensate analog black holes (10–100 ms). Simulations on a 48-qubit MERA lattice (bond dimension 8) confirm numerical robustness, and an observer-modified Ryu–Takayanagi prescription embeds the framework in AdS/CFT without requiring replica-wormhole or island constructions. By replacing ensemble-averaged Page curves with a causal, observer-specific mechanism, the model transforms the paradox from a bookkeeping puzzle into a falsifiable dynamical prediction. Here, S_max is the Bekenstein–Hawking entropy, γ(τ) is the modular-flow retrieval rate, and τ_char is a characteristic proper-time scale (geometric units c = G = 1).

Abstract: Background/Objectives: Hepatocellular carcinoma (HCC) is a leading cause of cancer-related mortality, with the Asia-Pacific (APAC) region bearing a disproportionate burden. This paper examines HCC challenges within seven APAC health systems, identifies key barriers at each stage of the patient journey, and proposes tailored, actionable solutions. To effectively address HCC challenges, a stepwise approach should prioritise high-impact solutions, focusing on prevention, early diagnosis, and expanding surveillance to maximise health outcomes and economic benefits, while tailoring strategies to each health system’s unique resources and constraints. Methods: A mixed-methods approach was used, including expert consultations from the 2024 HCC APAC Policy Forum, a literature review, and a review of Japan’s HCC management model. Data were collected through workshops and stakeholder feedback from healthcare professionals, policymakers, researchers and patient advocates across Australia, India, Malaysia, South Korea, Taiwan, Thailand, and Vietnam. Results: Key findings include significant disparities in HCC awareness, prevention, early detection, diagnosis, and access to treatment. Common challenges across APAC include limited public awareness, suboptimal surveillance infrastructure, and financial barriers to care. The integration of novel biomarkers and advanced surveillance modalities were identified as crucial priorities for improving early detection. Japan’s multi-faceted approach to HCC management serves as a successful model for the region. Conclusions: A customised and targeted approach is essential for reducing HCC burden across APAC. The proposed recommendations, tailored to each health system’s needs can significantly improve patient outcomes and reduce healthcare costs. Effective collaboration among stakeholders is necessary to drive these changes.

Abstract: This paper presents a Unified Field Theory based on the principle that all physical matter arises from the curvature of standing waves in time and space. Starting from the photon — a wave of pure frequency without mass — we show how mass emerges when resonance closes into a stable geometric loop. The electron is modelled as a time-locked standing wave formed by two interacting photon harmonics. The proton, by contrast, arises from full three-dimensional resonance, where orthogonal curvature locks time across all axes, where orthogonal curvature closes the loop across all space-time axesAt the heart of this model lies a resonance amplification constant derived from finite spiral growth. Using a physical form of the golden ratio, emerging naturally from Fibonacci dynamics, we demonstrate that mass is not intrinsic — it is the product of curvature closure in discrete, harmonic steps. Nature does not follow abstract infinities; it completes resonance through geometric necessity.This theory unifies particles, fields, atomic structures, and cosmic phenomena under one principle: curved standing waves in time-space. From photons to electrons, from proton geometry to cosmic expansion, all emerge from the same resonance logic. No external fields, no adjustable constants — only harmonic geometry shaping curved time.The universe does not play dice. It resonates.

Abstract: “Heretic” is commonly understood as perspectival term employed by insiders to describe a specific kind of outsider, namely, one who, not only holds to seriously objectionable positions, but having been one of us, should know better. Accordingly, ‘heretic religiosity’ denotes for members of the mother religion a dissenting form of religiosity that deviates too sharply from their own to be con-tained within it. Michael Walzer’s well-known idea of connected criticism, to which this paper’s title alludes, is a person who firmly opposes his community’s way of life, but choses to remain within the fold. The type of religious heresy I shall be looking at in the following pages not only choses to re-main within, but is contained within the fold. And not only that, but it is actually formative of the fold. But in that case, why should it be considered heretical?

Abstract: This cross-sectional study aimed to assess the prevalence of malnutrition and osteosarcopenia among elderly women with rheumatoid arthritis (RA) and explore their clinical impact. Sixty-five women over 65 years with RA were evaluated using GLIM criteria for malnutrition and EWGSOP2-based assessments for sarcopenia; bone status was measured by DXA, TBS, and 3D-DXA. Malnutrition was identified in 49.2% and osteosarcopenia in 52.3% of participants. A significant bidirectional association was observed: malnourished patients had higher rates of osteosarcopenia (65.6% vs. 34.4%; p< 0.05), and osteosarcopenic patients were more frequently malnourished (61.8% vs. 39.1%; p< 0.05). Both conditions were associated with older age, lower BMI, impaired muscle parameters, and reduced bone mineral density. Malnourished and osteosarcopenic patients reported worse fatigue and lower physical quality of life, despite similar inflammatory activity. Significant correlations were found between muscle mass indices and bone quality metrics assessed by 3D-DXA. These findings highlight a substantial burden of malnutrition and osteosarcopenia in elderly women with RA, even with well-controlled disease, reinforcing the need for integrated nutritional and musculoskeletal evaluation to prevent functional decline and frailty.

Abstract: Weeds present a pervasive challenge in agricultural fields. The integration of herbicide-resistant crops with chemical weed management offers an effective solution for sustainable weed control while reducing labor inputs, particularly in large-scale intensive farming systems. In this study, we found that a G311E mutant variant of the Arabidopsis MATE (multidrug and toxic compound extrusion) family transporter DTX6, designated DTX6m, confers robust resistance to bipyridyl herbicides in rice. DTX6m-overexpression lines exhibited marked resistance to paraquat and diquat, tolerating diquat concentrations up to 5 g/L, which is five-fold higher than the recommended field application dosage. Agronomic assessments demonstrated that grain yields of DTX6m-overexpression plants were statistically equivalent to those of wild-type plants. Moreover, the plants displayed beneficial phenotypic changes, such as accelerated flowering and a slight reduction in plant height. Seed morphometric analysis indicated that, in comparison with the wild-type control, DTX6m-transgenic lines exhibited altered grain dimensions while maintaining consistent 1000-grain weight. Nutritional assays further demonstrated that DTX6m increased the free amino acid levels in seeds, while normal protein and starch contents were retained. Collectively, these results establish that DTX6m effectively boosts rice resistance to bipyridyl herbicides, validating DTX6m as a candidate gene for engineering plant herbicide resistance and also implying a potential role for DTX6m in amino acid homeostasis in plants.

Abstract: The Inhibitor of DNA-binding (ID) proteins are dominant-negative regulators of basic helix-loop-helix (bHLH) transcription factors, playing a pivotal role in cellular differentiation, proliferation, and lineage plasticity. The Polycomb Repressive Complex 2 (PRC2), through trimethylation of histone H3 at lysine 27 (H3K27me3), mediates long-term transcriptional repression essential for development and cancer. Here, we hypothesize that ID proteins indirectly modulate PRC2 recruitment and function by altering chromatin accessibility and transcription factor occupancy, particularly during transitions between stem-like and differentiated states. We propose a mechanistic model whereby ID proteins prime the chromatin landscape for PRC2 activity, promoting a permissive environment for epigenetic silencing in stem/progenitor populations and tumor cells. This functional support could cause phenotypic plasticity in cancer and offer novel epigenetic targets for therapeutic intervention.

Abstract: Background: The motor response to human visual stimuli is unique and differs from the reaction to light-based visual stimuli. While laboratory-based tests offer valuable insights into athletes' basic perceptual-motor abilities, their translation to actual sports-specific tests is limited. Methods: Following a thorough warm-up, 44 collegiate-level male soccer players (age: 24.4 ± 2.5y, mass: 63.01 ± 7.3kg, Stature: 167.62 ± 6.3cm) from a tertiary institution completed the following tests: Sports Vision Test (SVT - 20-light proactive speed test), 40m sprint test (split times over 5, 10, and 20m), and a live Reactive Agility Test (RAT) entailing them to sprint, change direction (left or right in response to a live tester) and sprint again. Results: Numerous moderate correlations were seen between the RAT and various sprint distances (r &gt; 0.3, ES &gt; 0.3, P &lt; 0.05). The reaction speed relationship between the light-based (SVT) and live stimuli (RAT) test yielded a weak relationship (r &gt; 0.4, ES &gt; 0.5, P &lt; 0.05). Furthermore, hand-eye coordination speed did not predict acceleration or top speed, while the total RAT time did explain 10.5% of top speed (40m). Conclusions: Although limited, the correlations observed between the reaction to light test, sprint speed, and reaction to visual stimuli encourage the training thereof to maintain attentional focus and situational awareness, skills deemed crucial for reacting effectively in dynamic game situations. Future studies should aim to adapt contextual factors of laboratory-based tests to improve ecological validity to further explore the transferability of skills from the laboratory to the field.

Abstract: Background: Claustrophobia is defined as the fear of enclosed spaces, and it is a rather common specific phobia. Although the Claustrophobia Questionnaire (CLQ) is a valid questionnaire to measure claustrophobia, there are no studies to validate this tool in Greek. Thus, our aim was to translate and validate the CLQ in Greek. Methods: We applied the forward-backward translation method to translate the English CLQ in Greek. We conducted confirmatory factor analysis (CFA) to examine the two-factor model of the CLQ. We examined the convergent and divergent validity of the Greek CLQ by using the Fear Survey Schedule-III (FSS-III-CL), the NEO Five-Factor Inventory (NEO-FFI-NL-N), and the Spielberger’s State-Trait Anxiety Inventory (STAI). We calculated intraclass correlation coefficient (ICC) and Cronbach’s alpha to assess the reliability of the Greek CLQ. Results: Our CFA confirmed the two-factor model of the CLQ since all the model fit indices were very good. Standardized regression weights between the 26 items of the CLQ and the two factors ranged from 0.559 to 0.854. The convergent validity of the Greek CLQ was very good since it was correlated strongly with the FSS-III-CL and moderate with the NEO-FFI-NL-N and the STAI. Additionally, the Greek CLQ correlated higher with the FSS-III-CL than with the NEO-FFI-NL-N and the STAI indicating very good divergent validity. Reliability of the Greek CLQ was excellent since the ICC in test-retest study was 0.986 and the Cronbach’s alpha was 0.956. Conclusions: The Greek version of the CLQ is a reliable and valid tool to measure levels of claustrophobia among individuals.

Abstract: The synthesis method of the Pr-doped CeO2 catalyst support in Ni/Pr-CeO2 CO2 methanation catalysts is varied by changing the type/ basicity of the precipitating solution and the hydrothermal treatment temperature. The use of highly basic NaOH as the precipitating agent and elevated hydrothermal treatment temperature (100 or 180 oC) leads to the formation of structured Pr doped CeO2 nanorods and nanocubes, respectively, whereas the use of a mildly basic NH3 based buffer in the absence of hydrothermal treatment (i.e., co-precipitation) leads to an unstructured mesoporous morphology with medium-sized supported Ni nanoparticles. The latter catalyst (Ni/CP_NH3) displays a high surface area, high population of moderately strong basic sites, and favorable Ni dispersion. These properties lead to a higher catalytic activity for CO2 methanation (75% CO2 conversion and 99% CH4 selectivity at 350 oC) compared to the catalysts with structured nanorod and nanocube support morphologies, which are found to contain a significant amount of leftover Na from the synthesis procedure that can act as a catalyst inhibitor. In addition, the best performing Ni/CP_NH3 catalyst is shown to be highly stable, with minimal deactivation during time on stream operation.

Abstract: Despite these advances, several challenges persist, including measurement accuracy in unsupervised environments, the complexity of clinical data integration, and digital literacy gaps among older adults. Nevertheless, the integration of artificial intelligence (AI), predictive analytics, and remote rehabilitation tools is driving a shift toward more adaptive and individualized care models. This paper concludes that sensor-enhanced rehabilitation is no longer a future aspiration but an active transition toward a smarter, more accessible, and patient-centered paradigm in recovery after TKA.

Abstract: Prime numbers, long treated as isolated milestones on the number line, emerge here as the eigen-frequencies of a self-adjoint operator we call the Prime Laplacian. We prove—by Nelson’s commutator theorem, a profinite Fourier diagonalisation, and a compact-resolvent argument—that its spectrum is exactly the set of primes, each with multiplicity one and no continuous part. This “arithmetic drum” is then cooled: a Lorentzian heat kernel produces the trace formula Tr(e^-t T_Prime) = Σ_p e^-t p and a zeta-regularised determinant det' T_Prime ≈ 1.413. Embedding the same Lorentzian profile as an RG regulator yields an exact Wetterich flow, forging a bridge between prime spectra and functional renormalisation in quantum field theory. Sparse matrices up to size N = 4×10^2 confirm an N^-1 convergence rate toward prime eigenvalues, and open-source scripts push the numerics to N ~ 10^6. The framework extends naturally to twin-prime operators and speculative “zeta-resonance” Laplacians, hinting at fresh approaches to the Riemann Hypothesis and Planck-scale physics. In short, we recast primes as audible notes, supply the full sheet music, and invite both number theorists and quantum physicists to play along.

Abstract: The topic of forces on a dielectric slab between the two conductors of a parallel-plate capacitor is a classic textbook example that has drawn repeated attention of physicists for pedagogical reasons and beyond. It is pointed out in this paper that the critical role of Newton’s third law was not fully exposited in the current literature. This omission led to the current view that the force on the slab is solely due to the fringing field near the edges of a capacitor and, yet we can correctly compute the force while neglecting the fringing field. Considering the third law explicitly, multiple views are shown to be equally valid about where the forces are distributed. In a straightforward view presented in this paper, the force on the slab by the capacitor is mostly distributed deep inside the capacitor, more than near the edges. The textbook formula for the force is the zeroth-order approximation in terms of the fringing field. The inclusion of the fringing field will add correctional terms to the formula.

Abstract: This study investigates the effects of different biostimulants on the growth, yield, and stress

Abstract: Crack formation and rehabilitation of concrete elements, such as slabs and columns, in buildings located in areas with temperatures above 25 °C, where accelerated water evaporation significantly reduces their structural strength, was addressed. To mitigate this effect, an intelligent automated monitoring and curing system was developed, using integrated sensors and embedded controllers. The study was applied, experimental and explanatory, based on the hypothetico-deductive method. Work was carried out with physical models (specimens, columns and solid slabs) with 1:2:3 dosage, in which thermal and humidity sensors were integrated, and a visualisation system with an LCD screen was designed in Proteus for data collection. Quantitative analysis, with 95 % confidence, revealed a moderate and significant correlation (r = 0.587; p = 0.001) between the environmental thermo-hygrometer and the embedded DS18B20 sensor, showing effective heat transfer. The low correlation with the HD-38 sensor (r = 0.143; p = 0.468) indicated little influence of ambient humidity on internal humidity. The system restored the initial screams, used 1680 litres of water in 28 days, reducing consumption by 20 % compared to traditional methods. Higher evaporation was identified between 11:00 and 16:00 hours (UTC-5), a critical phenomenon not reported in previous studies in tropical areas.