Search Results (4,457)

In recent years, vision transformers have demonstrated remarkable superiority to CNNs in single image super-resolution (SISR), yet their heavy computational and storage costs hinder practical deployment. In this work, we revisit CNN-based residual learning for SISR and identify a specific yet overlooked problem: the residual signal is often biased by mapping errors during model training, and this bias can propagate and accumulate through layers. To this end, we revisit the commonly used residual learning and feature fusion in SISR and propose an adaptive residual correction network (AdaRCN) in this work. First, we introduce a residual correction mechanism that adaptively compensates for the bias in the residual, which is utilized to ease error accumulation and improve mapping accuracy. On the other hand, we generalize the standard identity shortcut to a weighted channel concatenation followed by a 1 × 1 convolution, which is a more versatile strategy for adaptive feature fusion. Our AdaRCN is built entirely upon a naive CNN without complex architecture design and training strategies, thus ensuring efficient inference and parallelization. Extensive experiments verify the benefits and effectiveness of residual correction and adaptive feature fusion in improving the representational capability of our model, enabling it to achieve impressive performance comparable to advanced SISR models with moderate overhead. Full article

The cost–performance relationship in mutual funds is a longstanding open question in financial economics, particularly when costs are assumed to exert a single, linear effect on returns. This study proposes an integrated causal machine learning framework to revisit this question using a panel of Hungarian open-ended public investment funds across all major asset classes—equity, bond, absolute yield, misc, money market, real estate, and commodity—covering 2017–2024. Six machine learning algorithms are benchmarked for return prediction, and Double Machine Learning, with fund-level cluster-robust inference and year fixed effects, is applied to estimate the effect of the Total Expense Ratio (TER) on next-year returns, under the identifying assumptions stated in the paper, while flexibly controlling for a set of observed fund-level confounders (size, NAV dynamics, volatility, past and cumulative performance, and fund age) without imposing a linear functional form. To move beyond average effects, a Causal Forest model—tuned using an out-of-fold, effect size-neutral selection criterion—estimates heterogeneous treatment effects across funds, and SHAP-based interpretation uncovers the mechanisms underlying this heterogeneity. The results show that, once the outcome is measured in the year following the one in which TER is observed and panel dependence is properly accounted for, the average TER effect is not robustly different from zero at the full-sample level; where a statistically robust effect emerges, it is negative rather than positive, concentrated in equity and absolute-yield funds, and largely confined to the period after 2022, which coincided with the war in Ukraine, rising interest rates, and heightened market volatility, although the research design does not identify which, if any, of these developments drove the change. Average-effect models are shown to conceal this heterogeneity, and the results are further shown to be sensitive to two methodological choices that might otherwise appear secondary—the timing convention linking cost and return, and the criterion used to select among competing heterogeneous-effects specifications—underscoring the importance of making such choices explicit. These findings demonstrate the added value of combining predictive and causal machine learning, together with identification-robust and panel-robust inference, for uncovering heterogeneity that conventional econometric approaches overlook and offer a transferable methodological template for causal machine learning applications in finance and other high-dimensional decision-making domains. Full article

Distributed offloading systems rely on periodic broadcasts to disseminate server state, yet propagation delays inevitably leave agents operating with stale information at decision time. Staleness is typically viewed as a performance limitation that should be minimized. This work revisits this assumption by studying its role in a distributed asynchronous offloading framework that operates under delayed and potentially stale load information. Players make independent server-selection decisions using locally available information and may optionally compensate for staleness through lightweight load-prediction mechanisms. A central finding is that staleness can act as an implicit coordination mechanism. By inducing heterogeneous and desynchronized perceptions of system state, it naturally diversifies agent decisions and mitigates collective migration oscillations that arise under perfect information sharing. These oscillations are shown to significantly delay convergence and can lead to unstable behavior in lightly loaded regimes. In contrast, stale yet diverse views prevent synchronized reactions and promote faster stabilization. The results further show that the value of prediction increases with communication staleness. As broadcast intervals grow and server-state information becomes increasingly outdated, prediction-based approaches substantially reduce performance degradation, improve load-balancing fairness, and lower migration activity relative to stale-information decisions. The numerical results show that increasing the broadcast interval from 0.5 s to 4 s causes a performance deterioration of approximately 95% for stale-information decisions, whereas prediction-based approaches limit this degradation to less than 15% over the same range. These findings suggest that the objective of distributed offloading should not be to eliminate staleness entirely, but rather to combine its stabilizing effects with lightweight predictive mechanisms that mitigate its negative impact on decision quality. Full article

Accurate assessment of nitrous oxide (N₂O) emissions from agricultural soils is essential for developing effective mitigation strategies. However, Pakistan’s current baseline estimates show inconsistencies. The updated Nationally Determined Contributions for 2018 report total greenhouse gas emissions from managed soils at 74.98 Mt CO₂e, while the first Biennial Update Report (BUR 1) estimates N₂O-specific emissions at 70.7 Mt CO₂e, rising under a business-as-usual scenario to 282.8 Mt CO₂e by 2030, an increase of nearly 300%. This study revisits these baselines using verified national datasets from Pakistan’s BUR 1 and the National Inventory Report 2021. The 2018 baseline was recalculated using the IPCC Tier 1 methodology, with fertilizer offtake and cropped-area data from national annual reports. Applying the updated IPCC AR6 global warming potential for N₂O (273) revised the BUR 1 2018 estimate to 62.26 Mt CO₂e, closely aligning with the NIR 2021 estimate of 62.4 Mt CO₂e. The study further disaggregates emissions by crop and source. Synthetic fertilizers were the largest direct source of N₂O emissions (14.72 Mt CO₂e), with wheat accounting for 50% (7.36 Mt CO₂e), mainly due to its extensive cropped area. Indirect emissions represented 35% of the revised baseline. Supported by uncertainty and sensitivity analyses, this refined baseline and its crop- and source-specific allocation provide a more reliable and transparent foundation for targeted mitigation in Pakistan’s most emission-intensive cropping systems. The findings also highlight the need for crop-specific fertilizer-use surveys to strengthen Tier 1 reporting where country-specific Tier 2 emission factors are unavailable. Full article

Phage therapy has been revisited as a biologically based strategy to tackle the escalating global crisis of multidrug-resistant (MDR) bacterial infections. Distinct from conventional antibiotics, bacteriophages target specific bacterial strains precisely, replicate locally at infection sites, penetrate bacterial biofilms, and exert synergistic effects with multiple antimicrobial agents. These inherent mechanistic advantages minimize collateral damage to the host’s commensal microbiota. However, existing regulatory frameworks—originally established for chemically synthesized, mass-produced drugs—fail to accommodate personalized, living biological phage products, leading to uncertain approval pathways and inconsistent manufacturing supervision. Clinical experience of phage therapy is predominantly derived from compassionate-use cases via multiple administration routes, including intravenous, inhaled, and topical delivery. This review systematically analyzes major challenges restricting clinical application, such as standardized production, quality control, pharmacokinetic characterization, rapid pathogen identification, and regulatory adaptation, as well as the limited performance of fixed phage cocktails against genetically heterogeneous bacterial populations. Current clinical practice demonstrates that phage therapy exhibits acceptable safety profiles across intravenous, inhaled, and topical administration routes, with promising therapeutic outcomes in otherwise untreatable MDR infections. Nevertheless, stable and reproducible clinical outcomes are hindered by multiple scientific and operational obstacles: the absence of unified standards for phage production and quality control, insufficient understanding of route-dependent pharmacokinetics, the imperative demand for rapid pathogen identification to enable precise phage matching, and the limited efficacy of fixed-cocktail regimens against genetically diverse clinical isolates. The successful integration of phage therapy into routine clinical practice relies on coordinated progress in diagnostic infrastructure construction, GMP-compliant phage repository establishment, international regulatory harmonization, and high-quality evidence generation through well-designed clinical trials. Rather than serving as a universal substitute for antibiotics, phage therapy is best implemented as a precision complementary component within comprehensive antimicrobial stewardship strategies. Full article

Joint-Embedding Predictive Architectures (JEPAs) have recently emerged as a promising paradigm for self-supervised representation learning by predicting latent embeddings from partial observations rather than reconstructing raw inputs. Although JEPA has demonstrated considerable success in computer vision and large-scale language models, the design of lightweight JEPA-inspired representation learning frameworks for natural language remains insufficiently understood. This paper proposes a lightweight span-masked JEPA-inspired framework that formulates language representation learning as latent semantic prediction from partially observed contextual inputs. The framework predicts target embeddings generated by a momentum-updated encoder while avoiding direct token reconstruction, thereby emphasizing semantic abstraction over lexical recovery. To investigate the influence of objective function design, cosine similarity and Variance–Invariance–Covariance Regularization (VICReg) are systematically compared under identical architectural and training conditions, with and without auxiliary masked language modeling (MLM) supervision. The proposed framework is evaluated on one controlled dataset and six benchmark datasets, including AG News, SST-2, IMDb, TREC, DBPedia, and Yelp Polarity. Experimental results consistently demonstrate that the cosine-based JEPA objective without auxiliary MLM supervision provides the best balance between representation quality, computational efficiency, and downstream linear probing performance. Statistical significance analysis using the Friedman test and pairwise Wilcoxon signed-rank tests with Holm correction further confirms that the cosine-based objective significantly outperforms the remaining configurations, while no significant difference is observed between VICReg with and without MLM supervision. The findings indicate that latent semantic prediction alone is sufficient for learning robust language representations under the proposed lightweight framework, whereas auxiliary token-level reconstruction does not consistently improve downstream performance despite increasing computational cost. Overall, this work provides reproducible empirical evidence and practical guidance for objective function design in lightweight JEPA-inspired language representation learning, offering an efficient experimental foundation for future self-supervised language models. Full article

Epithelial ovarian cancer (EOC) comprises biologically and immunologically distinct histotypes and frequently develops within a chronically inflamed tumor microenvironment, particularly in advanced disease with malignant ascites. This narrative review revisits translational observations from the 1990s–2000s showing impaired lymphomonocyte proliferation, altered Fas/CD25 signaling, and abundant cytokine release in ovarian cancer effusions. These findings are interpreted as direct evidence of dysfunctional immune activation, but not as retrospective proof of T-cell exhaustion according to contemporary molecular, transcriptional, epigenetic, or functional definitions. Modern single-cell and spatial studies provide independent evidence that malignant ascites is a dynamic ecosystem containing heterogeneous T-cell and macrophage states and that immune architecture differs across anatomical compartments and histotypes. Within this framework, cytokine signaling, macrophage plasticity, iron/redox biology, ferroptosis susceptibility, adipocyte–tumor crosstalk, and systemic metabolic dysfunction are considered at different levels of evidentiary strength, with ovarian cancer-specific data distinguished from pan-cancer or preclinical extrapolation. Clinical experience with immune-checkpoint inhibitors further illustrates the distinction between immune-cell presence and effective immunity: single-agent activity has generally been modest, whereas the phase III ENGOT-ov65/KEYNOTE-B96 trial demonstrates that clinically meaningful benefit can emerge in an appropriate therapeutic and biomarker-selected context. We propose that “inflammation without effective immunity” is best viewed as an overarching, histotype- and context-dependent immunometabolic framework rather than a uniform ovarian cancer phenotype. The concept remains hypothesis-generating and requires prospective validation before biomarker or therapeutic implementation. Full article

Humans and other great apes are unique among animals in lacking functional uricase, an enzyme that breaks down urate and allows its excretion. Modern-day humans have elevated serum urate, which is associated with various chronic health issues, including gout, metabolic syndrome, and hypertension. This prompts the question: what is the evolutionary benefit of the loss of uricase function, given that it may confer an increased risk of these chronic diseases? In this review, we explore numerous possible advantages, including some less commonly discussed, such as urate’s function as an antioxidant, its roles in ancestral survival, and its metabolic role in the context of our evolutionary history and the environmental pressures that have shaped human biology. We go on to examine the tradeoffs of urate’s evolutionary advantages, and how these features may be a double-edged sword for our health in our modern environment characterized by a caloric abundance of key nutrients, such as fructose, which are intimately linked with urate production and/or metabolism. Moving forward, we suggest that future work should explore the role of our evolutionary history in the rising prevalence of gout and other modern-day metabolic diseases. We further recommend that drugs and therapeutic strategies for diseases such as gout be informed by an understanding of the rich evolutionary context from which these maladies emerged and the unifying role of serum urate across these disease states. Full article

Generative data augmentation has been widely explored in Wi-Fi fingerprint-based indoor localization to reduce the cost of dense radiomap construction, with many studies reporting substantial localization improvements. However, these comparisons typically rely on default or weakly optimized baseline regressors, making it difficult to determine whether reported gains reflect genuine synthesis quality or merely compensate for suboptimal baselines. In this paper, we systematically investigate under which conditions generative augmentation is actually justified. We fine-tune four widely used localization regressors—kNN, SVR, XGBoost, and DNN—using Bayesian hyperparameter optimization and establish strong non-augmented baselines across radiomaps with controlled levels of spatial sparsity, constructed via farthest-point sampling. We then train five representative generative models—VAE, GAN, DDPM, DiT, and TDPM—within a unified augmentation pipeline that includes quality filtering and pseudo-labeling, and benchmark them against these baselines. Using two publicly available datasets, we show that none of the generative models consistently outperforms a non-augmented, fine-tuned baseline regressor such as XGBoost or kNN, across a wide range of sparsity levels. We further show that these conclusions are robust to three potential confounders: various proportions of synthetic data, the choice of localization regressor (ruling out circularity with the pseudo-labeling model), and the dataset itself, since the findings on the first dataset replicate on a second, structurally different building. These findings suggest that reported augmentation benefits in prior work may partly reflect under-optimized baselines rather than genuine synthesis quality, and that generative augmentation should be treated as a conditional last resort rather than a universal improvement strategy. Full article

Prenatal detection of agenesis of the corpus callosum (CC) anomalies is a challenge. Although a correct diagnosis of anomalies of the CC requires direct examination of the CC in the mid-sagittal plane of the foetal brain, the international guidelines on the mid-trimester morphology scan do not recommend such direct examination of the CC in low-risk populations because of the associated technical difficulties. Recently, such routine direct assessment has been recommended by several international experts in a consensus statement. The implementation of such routine direct assessment is not easy given the difficulties encountered in obtaining the mid-sagittal view of the CC. As such, it is the time to revisit the various two-dimensional and three-dimensional ultrasound techniques with a view to improve visualisation of the CC. The aim of this narrative review article is to discuss the use of various prenatal ultrasound screening methods of CC anomalies including standard axial views, more detailed axial views, the mid-sagittal view, the transvaginal approach, and 3D reconstruction. New insights on screening methods, and a pragmatic approach are also shared. Full article

Accurate channel state information (CSI) is essential for reliable orthogonal frequency division multiplexing (OFDM) transmissions, especially when training resources are limited and iterative receiver processing is employed. This paper revisits least squares (LS) and minimum mean square error (MMSE) channel estimation based on training sequences and analyzes their impact on an iterative turbo receiver framework. The initial channel estimate is obtained from an OFDM training transmission, while subsequent refinement is performed using soft information generated by a soft-input soft-output (SISO) equalizer and decoder. Unlike conventional approaches that keep the channel estimate fixed after the training phase, the proposed architecture enables decision-directed channel refinement using reconstructed transmit symbols. The OFDM stage is employed for channel estimation, whereas BER performance is evaluated using independently generated turbo-coded BPSK sequences transmitted through the analyzed channel. The performance analysis investigates the influence of training sequence length and signal-to-noise ratio (SNR) on iterative estimation gains. Simulation results show that, for short training sequences, the proposed iterative strategies can significantly improve BER performance compared with conventional non-iterative LS and MMSE estimators. These results provide practical insights for the design of training-efficient OFDM-based communication systems. Full article

This study aims to identify the key destination-related factors associated with tourists’ revisit behavior in peri-urban rural areas within Chinese metropolitan regions in the context of short-video marketing, and to reveal the structural relationships and hierarchical characteristics among these factors. First, the LDA topic model was employed to conduct text mining on authentic tourist-generated comments posted on Douyin, through which the core factors related to revisit behavior were identified and conceptually standardized based on tourists’ expressions. Building on this process, 153 experts in relevant fields were invited to evaluate the direction and strength of the relationships among these factors. An integrated DEMATEL–ISM–MICMAC approach was then applied to analyze their causal attributes, hierarchical structure, and systemic roles. The results indicate that the identified factors do not operate independently but instead form a multilayered structure with clear hierarchical characteristics. Among them, rural visual imagery, escape-oriented experience, rural lifestyle experience, and rural industry integration occupy deeper structural levels and exert relatively strong structural influences on factors located at intermediate and surface levels. The findings further suggest that the sustained attractiveness of rural tourism destinations in metropolitan regions cannot rely solely on short-video exposure or isolated “internet-famous” attractions; rather, it requires coordinated alignment among digital communication content, rural industries, lifestyle experiences, and tourism supply. By integrating tourist-generated content, natural language processing, and expert-based structural assessment, this study extends research on short-video tourism marketing from a systems perspective and provides practical insights for peri-urban rural destinations in Chinese metropolitan regions seeking to optimize the structural configuration of tourism resources, products and services, and marketing communication. Full article

This essay examines a long-term collaboration between writer and curator Dr George Vasey and the London-based organisation Intoart, which supports a group of 24 artists with learning disabilities. Throughout the collaboration, a broad array of practice-based curatorial methods have been employed—including field recording, creative writing, and drawing—to develop an interpretative framework for a growing organisational collection of over 6000 artworks. Drawing on John Dewey’s writing on art as experience, the essay expands on Dewey’s concept of thinking through doing in the context of disability arts, arguing for progressive studio provision as a site of exchange and community learning that emphasises an art of lived experience. The essay explores what an artist-led approach might mean in the context of disability, revisiting Dewey’s practice-led approach and foregrounding embodied and experimental forms of engagement. By centring the voice and intention of learning-disabled artists, the project focuses on facilitative strategies that merge the disciplines of art criticism, visual arts, curating, and pedagogy. The essay contributes to ongoing debates in curatorial and disability studies by demonstrating how facilitative and practice-led approaches can reshape interpretative frameworks, artist agency and knowledge production. Full article

Virtual reality (VR) applications are often designed as episodic experiences, with limited support for persistence, longitudinal revisitation, and structured integration of interaction data across sessions. This paper presents a persistent multi-user VR garden architecture that combines snapshot-based state restoration, structured event, movement, and transcript records, an asymmetric owner–visitor workflow, cloud-mediated speech transcription, and deferred AI-supported synthesis. The architecture separates the current spatial configuration of the environment from the interaction traces through which it evolves, supporting repeated access, state restoration, historical consultation, and post-hoc processing. A controlled technical validation using synthetic or researcher-generated inputs was conducted through Unity Editor/backend tests and on Meta Quest 3 hardware. Persistence was evaluated at 20, 100, and 250 objects in the Editor and at 1, 50, and 100 objects on Quest, with two Quest replicas per load. Application-level visitor restrictions were verified across nine prohibited write operations. A frozen production speech-to-text corpus completed 40/40 requests with a micro-averaged word error rate of 9.50% and a median end-to-end latency of 1.619 s. The deferred AI pipeline was additionally verified as a functioning technical integration, while limitations in semantic-detail preservation were observed. The results support the implementation-level feasibility of the proposed persistent and traceable VR architecture. The present evaluation does not establish backend-level authorization guarantees, general AI or NPC-grounding performance, user outcomes, or clinical effectiveness. Full article

Rotating machinery prognostics is essential for ensuring the reliability and operational safety of industrial systems. Although gated recurrent units (GRUs) have achieved competitive performance in remaining useful life (RUL) prediction, whether their internal dynamics are consistent with irreversible degradation mechanisms remains largely unexplored. This study revisits GRU-based prognostics from a physics-consistency perspective and analyzes the potential mismatch between gating mechanisms and degradation evolution. A full-life benchmarking framework is developed based on the XJTU-SY bearing run-to-failure dataset. A training-based health indicator (HI) is constructed through multi-domain vibration feature extraction and principal component analysis, where the degradation-state representation and RUL prediction objective are explicitly distinguished to avoid physically inconsistent supervision. Several representative approaches, including statistical models and deep learning architectures (LSTM, GRU, TCN, and Transformer), are evaluated using both prediction accuracy metrics (RMSE, MAE, and R²) and physical consistency criteria (monotonicity index, monotonicity violation index, and degradation trend consistency). Experimental results demonstrate that superior prediction accuracy does not necessarily guarantee physically consistent degradation modeling. Although GRU provides competitive RUL prediction performance, its hidden-state evolution and gating responses exhibit noticeable non-monotonic behaviors during degradation progression. These findings reveal a potential discrepancy between prediction-oriented recurrent learning mechanisms and irreversible degradation dynamics, highlighting the importance of incorporating physics-consistency evaluation into reliable data-driven prognostic models. Full article