Search Results (1,104)

Extreme summarization generates very short summaries, typically a single sentence, answering the question “What is the document about?”. Although large language models perform well in text generation, fine-tuning them for summarization often requires substantial computational resources that are unavailable to many researchers. In this study, we present an effective method for instruction-tuning open decoder-only large language models under limited GPU resources. The proposed approach combines parameter-efficient fine-tuning techniques, such as Low-Rank Adaptation (LoRA), with quantization to reduce memory requirements, enabling training on a single consumer-grade GPU. We fine-tuned a pre-trained decoder-only model on the XSum dataset using an instruction-following format. Experimental results demonstrate that the proposed decoder-only approach achieves competitive performance on the XSum dataset under strict GPU memory constraints. On the full test set, the proposed 2G–1R pipeline attains ROUGE-1/2/L F1 scores of 46.0/22.0/37.0 and a BERTScore F1 of 0.917, outperforming the individual generator models in lexical overlap and semantic similarity. Evaluation was conducted using traditional overlap-based metrics (ROUGE) and semantic metrics, including BERTScore and G-Eval. While remaining competitive in ROUGE compared to strong encoder–decoder baselines, the pipeline consistently produces summaries with higher semantic quality. These findings demonstrate that large decoder-only language models can be efficiently fine-tuned for extreme summarization on limited consumer-grade hardware without sacrificing output quality. Full article

This paper presents a novel high-order numerical method. The proposed scheme utilizes polynomial generating functions to achieve p order accuracy in time for the Grünwald–Letnikov fractional derivatives, while maintaining second-order spatial accuracy. By incorporating a short-memory principle, the method remains computationally efficient for long-time simulations. The authors rigorously analyze the stability of equilibrium points for the fractional vegetation–water model and perform a weakly nonlinear analysis to derive amplitude equations. Convergence analysis confirms the scheme’s consistency, stability, and convergence. Numerical simulations demonstrate the method’s effectiveness in exploring how different fractional derivative orders influence system dynamics and pattern formation, providing a robust tool for studying complex fractional systems in theoretical ecology. Full article

Ageing is a dynamic biological process involving interconnected physiological, psychological, and social changes, making the promotion of healthy ageing a global public health priority. The World Health Organization (WHO) defines healthy ageing as the process of developing and maintaining functional ability that enables well-being in older age. The WHO’s Decade of Healthy Aging (2021–2030) outlines four key action areas: changing attitudes toward ageing, creating age-friendly environments, delivering integrated and person-centred care, and ensuring access to long-term care. This Perspective examines yoga, a holistic mind–body practice integrating physical postures, breath regulation, and mindfulness, as a potentially safe, adaptable, and scalable intervention for older adults. Evidence suggests that yoga may improve flexibility, balance, mobility, and cardiovascular function, reduce pain, and support the management of chronic conditions commonly associated with ageing. Psychological and cognitive research further indicates reductions in stress, anxiety, and depressive symptoms, alongside potential benefits for attention, memory, and executive function. Improvements in health-related quality of life (HRQoL) have been reported across physical, psychological, and social domains, with benefits sustained through regular practice. Adaptations such as chair-based practices, restorative postures, and the use of props enhance accessibility and safety, allowing participation across diverse functional levels. Mindfulness and breath-focused components of yoga may additionally support emotional regulation, resilience, and psychological well-being, particularly among older adults experiencing stress or limited mobility. Yoga interventions are generally well tolerated, demonstrate high adherence, and can be delivered through in-person and digital formats, addressing common access barriers. Despite this growing evidence base, yoga remains underintegrated within health policy and care systems in the US, UK, and India. Strengthening its role may require coordinated efforts across research, policy, and implementation to support healthy ageing outcomes. Full article

Bacteriophages, traditionally viewed solely as antibacterial agents, are increasingly being studied for their immunomodulatory properties. In this study, we demonstrate that PM16 phage therapy not only effectively controls subcutaneous Proteus mirabilis infection in mice but also induces long-term specific humoral immunity against subsequent reinfection. This immunomodulatory effect was dose-dependent. In vitro, PM16 directly activates macrophages, leading to increased production of proinflammatory cytokines (tumor necrosis factor-α and interleukin-1β) and inducible nitric oxide synthase, and enhances macrophage bactericidal activity against P. mirabilis. We assume that the enhancement of the adaptive immune response is mediated not by the phage acting as a classical antigenic adjuvant but by its ability to prime innate immune cells, specifically macrophages. This priming leads to more efficient bacterial clearance, antigen presentation, and the formation of protective immunological memory. Full article

The formation and stability of river ice covers in regulated waterways are critical for uninterrupted hydro-electric operations. This study investigates the modelling of ice cover development in the Beauharnois Canal along the St. Lawrence River with the presence and absence of ice booms. Ice booms are deployed in this canal to promote the rapid formation of a stable ice cover during freezing events, minimizing disruptions to dam operations. Remote sensing data were used to assess the spatial extent and temporal evolution of an ice cover and to calibrate the river ice model RIVICE. The model was applied to simulate ice formation for the 2019–2020 ice season, first for the canal with a series of three ice booms and then rerun under a scenario without booms. Comparative analysis reveals that the presence of ice booms facilitates the development of a relatively thinner and more uniform ice cover. In contrast, the absence of booms leads to thicker ice accumulations and increased risk of ice jamming, which could impact water management and hydroelectric generation operations. Computational efficiencies of the RIVICE model were also sought. RIVICE was originally compiled with a Fortran 77 compiler, which restricted modern optimization techniques. Recompiling with NVFortran significantly improved performance through advanced instruction scheduling, cache management, and automatic loop analysis, even without explicit optimization flags. Enabling optimization further accelerated execution, albeit marginally, reducing redundant operations and memory traffic while preserving numerical integrity. Tests across varying ice cross-sectional spacings confirmed that NVFortran reduced runtimes by roughly an order of magnitude compared to the original model. A test GPU (Graphics Processing Unit) version was able to run the data interpolation routines on the GPU, but frequent data transfers between the CPU (Central Processing Unit) and GPU caused by shared memory blocks and fixed-size arrays made it slower than the original CPU version. Achieving efficient GPU execution would require substantial code restructuring to eliminate global states, adopt persistent data regions, and parallelize at higher level loops, or alternatively, rewriting in a GPU-friendly language to fully exploit modern architectures. Full article

High-precision prediction of the crystal diameter during the growth of electronic-grade silicon single crystals is a critical step for the fabrication of high-quality single crystals. However, the process features high-temperature operation, strong nonlinearities, significant time-delay dynamics, and external disturbances, which limit the accuracy of conventional mechanism-based models. In this study, mechanism-based models denote physics-informed heat-transfer and geometric models that relate heater power and pulling rate to diameter evolution. To address this challenge, this paper proposes a hybrid deep learning model combining a convolutional neural network (CNN), a bidirectional long short-term memory network (BiLSTM), and self-attention to improve diameter prediction during the shoulder-formation and constant-diameter stages. The proposed model leverages the CNN to extract localized spatial features from multi-source sensor data, employs the BiLSTM to capture temporal dependencies inherent to the crystal growth process, and utilizes the self-attention mechanism to dynamically highlight critical feature information, thereby substantially enhancing the model’s capacity to represent complex industrial operating conditions. Experiments on operational production data collected from an industrial Czochralski (Cz) furnace, model TDR-180, demonstrate improved prediction accuracy and robustness over mechanism-based and single data-driven baselines, supporting practical process control and production optimization. Full article

Interactions between alkaline solutions and the surface of pyrolytically coated graphite tubes (PCGTs) with/without a platform for determination of vanadium using high-resolution continuum source graphite furnace atomic absorption spectrometry (HR CS GFAAS) are discussed. Changes on the surface of tubes, lifetime of tubes, and formation of memory effect in the determination of vanadium in alkaline solutions (NaOH, Na₂CO₃, and real alkaline slag leachates) were investigated. Based on the results obtained, it is possible to state that HR CS GFAAS determination of vanadium content in alkaline solutions reveals that PCGTs with a platform are more susceptible than those without a platform to the formation of deposits and degradation of the platform surface, especially after the application of hydroxide environments. More marked and faster formation of deposits leads to shortening of the analytical lifetime of PCGTs with a platform (approx. 70 atomization/analytical cycles (ACs)) compared to PCGTs without a platform (approx. 290 ACs). The mechanical life of both types of tubes is comparable (approx. 500 ACs). Deposits formed on the internal surface of PCGTs can be removed in the presence of a carbonate environment and higher temperatures. Damage to the PCGT surface leads to the formation of scaled shapes and cavities, which can result in decreased absorbance due to losses of vanadium in the cavities (negative measurement error), or in increased absorbance by washing out of vanadium from the cavities (positive measurement error, and formation of memory effect). It was found that more frequent cleaning of PCGTs by performing ACs in an environment of 4 mol L⁻¹ HNO₃ can eliminate these unfavourable phenomena. Our results have shown that in the case of samples analysed with different sample environments (acidic vs. alkaline), the surface material of the tube/platform wears out more quickly, and therefore it is necessary to include a cleaning stage after changing the nature of the environment. Full article

This article examines how the Ahiska Turks—deported from Georgia’s Meskheti region to Central Asia in 1944—sustained their religious belonging under shifting Soviet and post-Soviet political and social conditions, and how this religious continuity became intertwined with processes of collective memory formation. Drawing on published archival materials, existing scholarship, and a long-term ethnographic corpus composed of fourteen life-history oral interviews conducted between 2006 and 2025 in Turkey and Kazakhstan, the study traces the multigenerational trajectories of ritual practice. The findings show that funeral ceremonies, mevlid gatherings, Ramadan practices, and domestic prayer circles function as “sites of memory” through which the trauma of displacement is reinterpreted and intergenerational belonging is continually reconstituted. These ritual forms generate a meaningful sense of continuity and communal resilience in the face of prolonged experiences of loss, uncertainty, and “placelessness.” Situated at the intersection of the anthropology of religion, cultural trauma theory, and Soviet/post-Soviet diaspora studies, the article conceptualizes Islam as more than a realm of belief: for the Ahiska Turks, it operates as a core cultural infrastructure that anchors post-displacement resilience, social organization, and collective memory. The study contributes to the literature by offering an integrated analytical framework that places the Ahiska community within broader debates on religion, memory, and forced migration; by examining rituals not only as emotional practices but also as institutional and cultural scaffolding; and by foregrounding the understudied post-traumatic religious experiences of Muslim diasporas. Full article

Floods are among the most destructive natural disasters, posing major challenges to human safety, property, and urban resilience. Effective communication of flood risk is therefore crucial for disaster preparedness and the sustainable management of smart cities. This study explores how interface design elements of flood hazard maps, including interaction modes and legend color schemes, influence users’ risk perception, decision support, and usability. An online questionnaire survey (N = 776) and a controlled 2 × 2 experiment (N = 40) were conducted to assess user comprehension, cognitive load, and behavioral responses when interacting with different visualization formats. Results show that slider-based interaction significantly reduces task completion and map-reading times compared with drop-down menus, enhancing usability and information efficiency. Multicolor legends, although requiring higher cognitive effort, improve users’ risk perception, engagement, and memory of flood-related information. These findings suggest that integrating cognitive principles into interactive design can enhance the effectiveness of digital disaster communication tools. By combining human–computer interaction, visual cognition, and smart governance, this study provides evidence-based design strategies for developing intelligent and user-centered flood hazard mapping systems. The proposed framework contributes to the advancement of smart urban resilience and supports the broader goal of building safer and more sustainable cities. Full article

Complex landslides have characteristics and parameters that are difficult to analyze. The landslide on 16 June 2024 in the rural community of Quilloturo (Tungurahua, Ecuador) caused severe damage (14 deaths, 24 injuries, and hundreds of affected families) related to the area’s geological, social, and anthropogenic conditions. Its location in the eastern foothills of Ecuador’s Cordillera Real exacerbated the effects of a landslide involving various processes (mud and debris flows, landslides, and rock falls). This event was preceded by intense rainfall lasting more than 10 h, which accumulated and caused natural damming of the streams prior to the event. The lithology of the investigated area includes deformed metamorphic and intrusive rocks overlain by superficial clayey colluvial deposits. The relationship between the geological structures found, such as fractures, joints, schistosity, and shear, favored the formation of blocks within the flow, making mass movement more complex. Geomorphologically, the area features a relief with steep slopes, where ancient landslides or material movements, composed of these colluvial deposits, have already occurred. At the foot of these steep slopes, on plains less than 300 m wide and bordered by the Pastaza River, there are human settlements with less than 60 years of emplacement and a complex history of territorial occupation, characterized by a lack of planning and organization. The memory of the inhabitants identified mass movements that have occurred since the mid-20th century, with the highest frequency of occurrence recorded in the last decade of the present century (2018, 2022, and 2024). Furthermore, it was possible to identify several factors within the knowledge of the inhabitants that can be considered premonitory of a mass movement, specifically a flood, and that must be incorporated as critical elements in decision-making, both individual and collective, for the evacuation of the area. Full article

Alzheimer’s disease (AD) is a progressive neurodegenerative disorder characterized by memory loss, cognitive decline, and oxidative stress-driven neuronal damage. Catalase, a key antioxidant enzyme, plays a vital role in decomposing hydrogen peroxide (H₂O₂) into water and oxygen, thereby protecting neurons from reactive oxygen species (ROS)-mediated toxicity. In AD, the catalase function is compromised due to reduced enzymatic activity and aggregation, which not only diminishes its protective role but also contributes to amyloid plaque formation through catalase-Aβ co-oligomers. Hence, therapeutic strategies aimed at simultaneously preventing catalase aggregation and enhancing its enzymatic function are of great interest. In this study, we screened twelve naturally occurring metabolites for their ability to modulate catalase aggregation and activity. Among these, dimethylglycine (DMG) emerged as the most potent candidate. DMG significantly inhibited thermally induced aggregation of catalase and markedly enhanced its enzymatic activity in a concentration-dependent manner. Biophysical analyses revealed that DMG stabilizes catalase by promoting its native folded conformation, as evidenced by increased melting temperature (T_m), higher Gibbs free energy of unfolding (ΔG°), and reduced exposure of hydrophobic residues. TEM imaging and Thioflavin T assays further confirmed that DMG prevented amyloid-like fibril formation. Molecular docking and dynamics simulations indicated that DMG binds to an allosteric site on catalase, providing a structural basis for its dual role in stabilization and activation. These findings highlight DMG as a promising therapeutic molecule for restoring catalase function and mitigating oxidative stress in AD. By maintaining catalase stability and activity, DMG offers potential for slowing AD progression. Full article

Background: Electrochemotherapy (ECT) is a clinically validated local ablative treatment increasingly recognized for its ability to induce immunogenic cell death and stimulate antitumor immunity. Its combination with immune checkpoint inhibitors, such as anti-PD-1 antibodies, may enhance systemic immune responses and improve therapeutic efficacy, particularly in poorly immunogenic tumors. Methods: We evaluated the antitumor effectiveness of ECT combined with a murine analog of the anti-PD-1 antibody in four syngeneic murine tumor models with differing histology and immune status: WEHI fibrosarcoma, CT26 and MC38 colorectal carcinoma, and 4T1 mammary carcinoma. In vitro cytotoxicity assays assessed tumor cell sensitivity to ECT, while in vivo experiments evaluated complete response (CR) rates, immune cell infiltration, and long-term immune memory through secondary tumor challenge. Immunohistochemical analysis of CD4⁺, CD8⁺, and granzyme B⁺ effector cells. Results: In vitro, WEHI cells exhibited the highest sensitivity to ECT. In vivo, ECT monotherapy induced CRs in 100% of WEHI tumors, 60% of CT26, 17% of 4T1, and 15% of MC38. The addition of anti-PD-1 significantly enhanced outcomes in less responsive models, increasing CRs to 90% in CT26, 91% in MC38, and 53% in 4T1. Combination therapy promoted pronounced infiltration of CD4⁺, CD8⁺, and granzyme B⁺ T cells and the formation of tertiary lymphoid structure, particularly in MC38 tumors. Secondary challenge experiments confirmed long-term immune memory in CT26 and MC38 models and induced memory in 4T1, which was absent following monotherapy. Conclusions: ECT synergizes with PD-1 blockade to potentiate local and systemic antitumor immunity, overcoming immune resistance in poorly immunogenic tumors. These findings support further clinical development of ECT in combination with immune checkpoint inhibitors as a component of personalized cancer immunotherapy. Full article

Background: Resveratrol (3,5,4′-trihydroxy-trans-stilbene, RVT) is one of the most extensively studied natural polyphenols, with numerous health benefits documented in the literature. One of its most characterized biological properties is the strong antioxidant capacity. However, its poor biopharmaceutical properties limit its in vivo applicability. In this study, we conducted a detailed comparative analysis of the antioxidant and protective capacity of pure and loaded into Pluronic micelles resveratrol. Methods: Various in vitro antioxidant assays, such as DPPH, ABTS, superoxide anion radical scavenging, ferric (FRAP), and copper-reducing power assay (CUPPRAC), and iron-induced lipid peroxidation were performed. In addition, the in vitro 6-OHDA model of neurotoxicity in brain synaptosomes and the in vivo scopolamine (Sco)-induced model of cognitive impairment in rats were also employed. The main antioxidant biomarkers—the levels of lipid peroxidation (LPO) and total glutathione (GSH), as well as activities of superoxide dismutase, catalase, and glutathione peroxidase—were measured in the cortex and hippocampus. Results: The results from the in vitro tests demonstrated better ferric-reducing power activity and better neuroprotective capacity of the micellar resveratrol (mRVT), as evidenced by preserved synaptosomal viability and maintained GSH levels in a concentration-dependent manner in 6-OHDA-induced neurotoxicity. Regarding the in vivo results, mRVT (10 µM concentration) was the most effective treatment in supporting recognition memory formation in dementia rats. Further, mRVT demonstrated better LPO protective capacity in the hippocampus and GSH preserving activity in the cortex than the pure drug. Conclusions: The incorporation of resveratrol in polymeric micelles could enhance its antioxidant and neuroprotective effects. Full article

With the advancement of petroleum engineering, the increasing complexity of formations and unpredictable conditions make wellbore pressure prediction more challenging. Accurate bottom-hole pressure (BHP) prediction is crucial for the safe and stable development of oil and gas reservoirs. Solving the partial differential equations (PDEs) governing fluid flow is key to this prediction. As deep learning becomes widespread in scientific and engineering applications, physics-informed neural networks (PINNs) have emerged as powerful tools for solving PDEs. However, traditional PINNs face challenges such as insufficient fitting accuracy, large errors, and gradient explosion. This study introduces MetaPress, a novel physics-informed neural network structure, to address inaccurate formation pressure prediction. MetaPress incorporates a meta-learning-based embedding function that integrates spatial information into the input and forget gates of Long Short-Term Memory networks. This enables the model to capture complex spatiotemporal features of flow problems, improving its generalization and nonlinear modeling capabilities. Using the MetaPress architecture, we predicted BHP under single-phase flow conditions, achieving an error of less than 2% for L2. This approach offers a novel method for solving seepage equations and predicting BHP, providing new insights for subsequent studies on reservoir fluid flow processes. Full article

Vaccination remains a cornerstone of livestock disease control, yet its effectiveness under field conditions is often compromised by concurrent infections, particularly parasitic helminths. This review explores how infections shape vaccine-induced immunity in cattle, emphasizing the immunoregulatory mechanisms by which helminths interfere with protective responses. Chronic infections with Fasciola hepatica and Ostertagia ostertagi induce Th2-biased and regulatory immune environments that suppress antigen presentation, cytokine production, and memory formation and maintenance, leading to reduced vaccine efficacy. Evidence from experimental and field studies is scarce and constitutes a gap in our knowledge on how vaccines work in the field. Available data indicate that infection timing, intensity, and chronicity critically determine the extent of vaccine interference. The review highlights diagnostic approaches that can support targeted deworming before vaccination and proposes integrated management strategies combining parasite control, immunization, and nutritional optimization. Such approaches can mitigate helminth-driven immune suppression, enhance herd protection, and reduce dependence on anthelmintics. However, the impact of helminth infections on vaccine efficacy in cattle should be further assessed in the field. Understanding parasite–vaccine interactions is essential to refine vaccination programs, guide the development of next-generation vaccines, and promote sustainable livestock health in parasite-endemic areas. Full article