Search Results (336)

The operation and management of urban dynamic systems often rely on information, including holiday and event calendars, to anticipate deviations from routine demand. However, such calendars cannot capture irregularly announced holidays or undocumented physical disruptions and require continuous maintenance. This study proposes Structural Equilibrium, a reference state for interpreting urban dynamic systems without external calendars. Temporal variation is represented as Concentration–Dispersion–Relative Balance (CDR) structural states under two representative Structural Equilibria: Information-oriented and Stability-oriented Equilibria. Applying the two Equilibria to Jeju Airport electricity consumption and Gangnam Station subway passenger-flow data showed that structural representations diverged at specific times, producing structural Gaps. Moving-block bootstrap analysis showed that these divergences exceeded the range typically expected under the dataset’s own temporal dependence structure. Without calendar information, the divergences corresponded to the 2024 Chuseok holiday period in the subway data and a government-designated temporary public holiday and major snowstorm in the airport data. One high-Gap observation at Jeju Airport had no external cause despite an unremarkable raw magnitude, showing that structural analysis can reveal unusual conditions overlooked by magnitude-based monitoring. These findings indicate that Structural Equilibrium supports calendar-independent interpretation and adaptive monitoring under changing conditions, contributing to resilient and sustainable urban infrastructure operation. Full article

Background: Breast cancer continues to be a major contributor to illness among women in Jordan, particularly in underserved regions. Despite the establishment of mammography services in southern Jordan, no formal quality assurance audit has evaluated diagnostic performance or geographic equity of access. This study aimed to evaluate the diagnostic performance of mammography services using established quality assurance indicators and to investigate geographic variation in mammographic findings and diagnostic pathways among women attending a regional referral center in southern Jordan. Methods: This study employed a retrospective hospital-based clinical audit that was conducted using mammography records from Ma’an Hospital between 2016 and 2024. A total of 592 women aged 20 years and older who underwent mammography were included. Data extracted from electronic and paper-based medical records included demographic characteristics, mammography indication, BI-RADS^® classifications, follow-up procedures, reporting timelines, and histopathological outcomes. Histopathological confirmation served as the reference standard for diagnostic accuracy analyses. Diagnostic performance was evaluated using Cancer Detection Rate (CDR), sensitivity, specificity, positive predictive value (PPV), and negative predictive value (NPV). Geographic variation was evaluated according to place of residence, while multivariable logistic regression was used to determine whether associations remained independent of age and mammography indication. Results: Of the 592 mammography examinations, 354 (59.8%) were performed for screening purposes and 238 (40.2%) for diagnostic evaluation. BI-RADS^® 1 and BI-RADS^® 2 were the most frequently assigned categories, accounting for 38.3% and 28.7% of examinations, respectively, whereas BI-RADS^® 4 and 5 findings accounted for 6.3% and 1.0%. The overall CDR was 52.4 per 1000 examinations with sensitivity (93.9% (95% CI 80.4–98.3%)) and specificity (97.4% (95% CI 95.3–98.5%)), positive predictive value (PPV) (73.8% (95% CI 58.0–86.1%)), and negative predictive value (NPV) (99.5% (95% CI 98.2–99.9%)). Suspicious findings (BI-RADS^® 4–5) were significantly more frequent among urban than rural women (9.4% vs. 3.7%; adjusted OR 3.05, 95% CI 1.37–6.78), whereas reporting delay, ultrasound use, biopsy rate, and further follow-up did not differ significantly by residence. Conclusions: Mammography services at Ma’an Hospital exhibit high diagnostic accuracy and strong clinical performance in breast cancer detection over the nine-year study period. Women in rural areas were less likely to have suspicious mammographic findings detected than those in urban areas, regardless of age and mammography type. No significant differences were observed in reporting timelines or follow-up procedures by place of residence. These findings underscore the value of regional quality assurance audits in identifying service disparities and guiding improvements in breast cancer screening and diagnostic services, especially in underserved communities. Full article

Water security is a significant challenge for the Saudi Arabia Kingdom’s development and stability, affecting other economic sectors beyond the water sector. Insufficient water resources are causing economic and social crises; addressing this issue is crucial for the country’s growth and stability. This study aims to manage water resources in central Saudi Arabia (Wadi AS SAHBAA) through a two-level approach. The first level involves extracting rainfall amounts from satellite imagery to predict future rainfall intensity using PERSIANN-CCS-CDR data. The second level focuses on monitoring groundwater recharge using geographic information systems (GIS) and remote sensing techniques. The study further seeks to understand rainstorm behavior influenced by climate variability and applies geomatics techniques for quantitative analysis. The results show that the Wadi AS SAHBAA basin receives an annual precipitation of 116.6 mm/year and mean annual precipitation of 9.7 mm. The year 2019 experienced the highest recorded precipitation, reaching 222.3 mm and mean annual precipitation of 18.5 mm. Between 2013 and 2019, the AS SAHBAA region experienced increased rainfall driven by intense storm events; however, it declined during the period 2020–2022. Moreover, 2021 recorded the lowest annual precipitation of 53.8 mm with mean annual precipitation of (4.5 mm), possibly linked to reduced storm activity due to the COVID-19 pandemic. The study uses several methods to estimate groundwater recharge from rainfall and concludes that the average infiltration during 2013–2022 was varying from 1.59–36.63 mm, representing about between 1.03 and 28.5% of total rainfall. This is reflected in groundwater storage capacity, which ranges from approximately 1.65 to 38.27 million m³ yearly. This study provides a framework for monitoring precipitation and groundwater recharge and offers practical recommendations for regional development and sustainable water management. Full article

Functional cure of chronic hepatitis B virus (HBV) infection remains a significant challenge, making viral-entry-blocking antibodies a promising antiviral strategy. Here, we developed a structure-guided computational workflow for affinity-enhancing candidate mutations at the interface between the humanized neutralizing antibody HzKR127 and the HBV preS1 peptide epitope. Based on the crystal structure of the HzKR127–preS1 complex, we performed single-site saturation mutagenesis across the paratope, evaluating variants with a consensus effect score integrated from seven computational models. Benchmarking against published alanine scanning data showed that our consensus score effectively identified major-affinity-loss residues, achieving ROC AUC values of 0.81 and 0.84 for residue-level and site-level predictions, respectively. Mutational profiling revealed distinct asymmetric mutational responses, with broad intolerance on the preS1 side and localized favorable substitutions within antibody CDRs. Multilevel prioritization identified 26 antibody-side candidates, 17 of which showed improved HADDOCK refinement scores compared to the wild type. In particular, the H:D97W/F/Y substitutions presented the strongest structural rationale for enhancing improved interfacial packing through aromatic hydrophobic contacts with preS1 Phe10. These findings provide a prioritized list of candidates for experimental validation and a practical framework for the rational optimization of antibodies targeting functionally constrained viral epitopes. Full article

Accurate estimation of Alzheimer’s disease (AD) severity from structural magnetic resonance imaging (MRI) remains difficult, as disease-associated anatomical alterations are often subtle and publicly available datasets are typically too small to support robust deep learning model training. This feasibility study sought to determine how much Alzheimer’s disease spectrum-related information could be extracted from a small structural MRI cohort using a deliberately lightweight two-dimensional convolutional neural network (2D CNN), and whether transfer learning improves model performance. This study was intended as a methodological proof of concept rather than the development of a clinically deployable diagnostic tool. Structural scans and Clinical Dementia Rating (CDR) labels from the OASIS-1 dataset were filtered to 214 subjects: 124 cognitively normal (CN), 65 with mild cognitive impairment (MCI; CDR = 0.5), and 25 with AD-level impairment (CDR ≥ 1). A compact 2D CNN trained from scratch and a transfer learning model (frozen ImageNet MobileNetV2 features) were evaluated on four binary tasks (CN vs. AD, MCI vs. AD, CN vs. MCI, and CN vs. any impairment) under identical pre-processing and subject-level repeated 5-fold cross-validation (10 repeats), with the decision threshold tuned only on an inner split. Discrimination was summarized by ROC-AUC with 95% confidence intervals (CIs), permutation tests against chance, and per-task sensitivity and specificity. The from-scratch CNN recovered only a broad normal-versus-impaired signal (CN vs. any impairment AUC 0.59) and was at chance on adjacent-stage tasks (MCI vs. AD 0.41; CN vs. MCI 0.51). Transfer learning improved every task: CN vs. AD AUC 0.745 (95% CI 0.730–0.763), CN vs. any impairment 0.642, CN vs. MCI 0.601, and MCI vs. AD 0.599. On an independent OASIS-2 cohort, the transfer learning CN vs. AD model retained AUC 0.748. In this small-data regime, transfer learning recovers substantially more Alzheimer-spectrum signals than a from-scratch CNN, but performance remains modest because it is bounded by CDR-based, non-biomarker-confirmed labels, suggesting the model separates CDR-defined cognitive-status groups rather than detecting AD pathology. Full article

Background: The relationship between general anesthesia (GA) exposure and cognitive function remains controversial. This study aimed to investigate the association between GA exposure and cognitive severity, as measured using the Clinical Dementia Rating Sum of Boxes (CDR-SB) and Global Deterioration Scale (GDS), across different diagnostic groups. Methods: We used a de-identified dataset from the Korea Dementia Research Center’s Trial Ready Registry (KDRC TRR). Participants were classified as having normal cognition, mild cognitive impairment (MCI), or dementia. Logistic regression analyses were performed to identify the variables associated with CDR-SB and GDS scores within each group. Linear regression analyses were also performed for the amyloid positron emission tomography standardized uptake value ratios (SUVRs). Results: Among 688 participants, 258 were classified as having normal cognition, 245 as MCI, and 185 as dementia. In the normal group, GA exposure was associated with higher CDR-SB (OR = 2.33 [95% CI: 1.06–5.08, p = 0.032]) and GDS scores (OR = 2.29 [95% CI: 1.16–4.55, p = 0.017]). In addition, in this group, GA exposure was significantly associated with higher SUVRs in multivariable analyses (β = 0.09, 95% CI: 0.004–0.18, p = 0.040). No consistent associations were observed in the MCI or dementia groups. Conclusions: GA exposure after the age of 50 was associated with greater cognitive severity and higher amyloid burden in cognitively normal individuals. Prospective longitudinal studies incorporating detailed anesthetic, surgical, and perioperative data are required to clarify the nature of this association. Full article

Evaluating joint predictive performance for multivariate hydroclimatic models requires metrics that simultaneously assess marginal accuracy and cross-variable dependence recovery. Existing metricsthe Energy Score, Variogram Score, and their derivativesdo not adapt to the structural complexity of the residual correlation matrix, treating a single correlated pair identically to a fully dense dependence structure. We propose two novel metric families: Metric E (E-CVWMD: Enhanced Coefficient-of-Variation Weighted Marginal-Dependence) and Metric E2 (E-CVWMD-Pairwise), which are designed for mixed-type multivariate responses combining continuous and binary outcomes within a cross-validation framework. We position Metrics E and E2 as diagnostic ranking tools for comparing competing models rather than as strictly proper scoring rules, and we provide a strictly proper Log-Loss variant (E-LL/E2-LL) for applications that require the full properness guarantee. Metric E assigns variable-level weights proportional to the coefficient of variation (CV) of each outcome on the training partition and adaptively calibrates the marginal-dependence trade-off parameter ${\alpha }^{\ast }$ via a global distance-correlation test. Metric E2 refines this by replacing the global test with a pairwise Spearman screening index $\widehat{\pi }$, the proportion of variable pairs with significant residual correlationwhich maps linearly to ${\alpha }^{\ast }\left(\widehat{\pi }\right)=1-\widehat{\pi }/2\in [0.5,1]$. Applied to the validation of a Generalized Multivariate Functional Additive Mixed Model (GMFAMM) on 62 Valle del Cauca meteorological stations ($N_{\mathrm{test}}\approx$ 31,663), the naive significance-based index saturates ($\widehat{\pi }=1.0$) at this large sample sizeevery pair, including correlations as small as $|{\widehat{\rho }}_s| \approx 0.01$, is flagged “significant”which is precisely the sample-size sensitivity we address. Under the effect-size screening ($|{\widehat{\rho }}_s| \ge 0.05$), three negligibly correlated pairs are excluded, yielding $\widehat{\pi }=0.70$ and ${\alpha }_{E2}^{\ast }=0.65$, a better-calibrated weight than Metric E’s ${\alpha }_E^{\ast }\approx 0.797$ under the same data. A large-scale simulation study with 37,440 model evaluations confirms that Metric E inverts the correct ranking at correlation levels $\rho \ge 0.40$ (CDR = 0%), while E2 maintains correct discrimination in 14 of 15 simulation conditions (M1 vs. M3). We also delimit the metrics’ scope: E2 degrades under near-saturated uniform dependencea regime in which the strictly proper Energy Score remains preferableand the pairwise index is sensitive to sample size, for which we provide an effect-size-based variant. An R package (mvmetrics v0.2.0) implementing both metrics, the Log-Loss variant, alternative weighting schemes, and the effect-size screening is publicly available. Full article

The Philippines achieved World Health Organization (WHO) certification for the elimination of leprosy as a public health problem in 1998. Despite this milestone, new cases continue to be reported each year, highlighting the need for sustained surveillance and interventions to achieve zero transmission. This paper provides an update on the country’s progress toward interruption of leprosy transmission using national surveillance data and programmatic reports from 2020–2024. Quantitative data were obtained from the Department of Health (DOH) Field Health Services Information System (FHSIS), while policy and programmatic information were drawn from national reports, WHO guidance, and implementation reviews. Descriptive analyses were conducted to examine trends in prevalence, case detection rates (CDRs), and age-sex distribution patterns to identify high-risk groups. Leprosy prevalence declined from 0.41 per 10,000 population in 2020 to 0.11 in 2024, remaining below the WHO elimination threshold. The CDR increased from 0.45 in 2022 to 1.17 per 100,000 in 2024, indicating recovery of active surveillance after COVID-19-related disruptions. Most newly detected cases occurred among adults aged 20–59 years (72%), although continued detection among children aged 0–14 years (6–7%) suggests ongoing transmission in selected endemic areas. Key program strengths include policy integration and WHO-supported surveillance initiatives, while major barriers include stigma, uneven local implementation, and limited access to rehabilitation. The Philippines has maintained low national prevalence while strengthening efforts toward transmission interruption. Continued investment in surveillance, contact tracing, stigma reduction, and integrated neglected tropical disease (NTD) services will be essential to achieving zero transmission, zero disability, and zero discrimination by 2030. Full article

Decision-making in cyber defense exercises (CDX) is shaped by technical, emotional, motivational, and collaborative human factors under uncertainty and time pressure. This study proposes a human-centered Adaptive Neuro-Fuzzy Inference System (ANFIS) framework to model and predict Counterfactual Decision Reflection (CDR) outcomes in CDX environments. Two complementary datasets representing technical, emotional, motivational, and teamwork-related dimensions were collected from the international Lithuanian Armed Forces cyber defense exercise Amber Mist 2024 and analyzed using Spearman correlation, 3D regression surface modeling, fuzzy rule extraction, and ANFIS prediction to investigate the relationship between human factors and CDR. The results demonstrated that teamwork, communication, and collaboration have a stronger influence on decision stability than isolated technical competencies. Baseline ANFIS evaluation indicated that triangular membership functions provided the best generalization, while generalized bell functions achieved the lowest training errors. To improve model robustness, multiple synthetic data augmentation methods were evaluated. The augmented ANFIS models substantially improved predictive performance, reducing testing error values significantly. The findings confirm that synthetic-data-enhanced neuro-fuzzy modeling provides an effective and interpretable framework for analyzing human-centered cybersecurity decision-making processes in cyber defense exercises. Full article

Atmospheric CO₂ concentrations have reached significant levels during the industrial era, necessitating the implementation of effective carbon dioxide removal (CDR) technologies. Enhanced Rock Weathering (ERW) using basalt has emerged as a high-potential strategy, leveraging its mafic composition to sequester CO₂ as stable carbonates. This review analyzes ERW’s geochemical processes, application methods, and multifaceted co-benefits, such as restoring “background fertility” and improving soil structure. The literature indicates that while small-scale applications range from 1.5 to 6 Mg·ha⁻¹·yr⁻¹, intensive agricultural rates typically reach 40–100 Mg·ha⁻¹·yr⁻¹. Global models estimate a sequestration potential of up to 4.9 × 10⁹ Mg CO₂·yr⁻¹ for basalt, although field-scale results vary significantly, reaching uptake rates of up to 4 Mg CO₂·ha⁻¹ depending on pedological conditions and crop types. Despite this promise, transitioning to large-scale deployment faces critical hurdles, including operational difficulties in mechanized spreading and a scarcity of audited, long-term field data. Future research must prioritize standardized protocols and comprehensive economic analyses to bridge the gap between theoretical models and empirical evidence. Ultimately, ERW represents a multifaceted solution for climate stabilization and sustainable food security, provided that sequestration efficacy and environmental safety are rigorously verified through high-application field trials. Full article

Candidozyma auris (formerly known as Candida auris) has emerged as a formidable clinical fungal pathogen as a result of its multidrug resistance and persistent colonization capabilities. In this study, three clinical C. auris strains (namely C. auris strain 01, C. auris strain 03, and C. auris strain 13) with distinct origins were characterized to investigate their phenotypic variations and mechanisms of azole resistance. Comprehensive profiling revealed significant inter-strain differences in biofilm formation, cell surface hydrophobicity, adhesion capacity, and phospholipase activity. Testing for antifungal susceptibility showed that the three clinical strains exhibited different minimum inhibitory concentrations for multiple azoles (fluconazole, voriconazole, and itraconazole) and echinocandins (anidulafungin and micafungin). Sequencing identified Y132F mutations in the ERG11 gene of the three clinical strains. Mechanistic investigations demonstrated that fluconazole exposure significantly upregulated the expression of efflux pump genes (CDR1 and CDR2) and the genes encoding their transcriptional regulators (MDR1 and TAC1b). In a murine skin colonization model, comparing data from the standard strain C. auris strain CBS12766 and clinical strains of C. auris strain 03 and C. auris strain 13 exhibited a significantly higher fungal burden of tissue, whereas strain C. auris strain 01 showed an intermediate level. Host immunity response analysis revealed that expression of the IL-1β gene was significantly elevated in C. auris strain CBS12766-infected mice, while expression of IL-6 and CXCL-1 genes was predominantly increased in the C. auris strain 01, with TNF-α gene expression levels being comparable across all strains. Histopathological examination confirmed local infiltration of inflammatory cells and mild epidermal edema, indicating active host immune engagement. Overall, our findings highlighted substantial phenotypic heterogeneity, different colonization capacities, and differences in expression of inflammatory cytokines among the C. auris strains. Further investigations into fluconazole-response mechanisms identified enhanced efflux pump activity, along with ERG11 gene Y132F mutations and transcription factor modulation among these clinical strains. Full article

Background and Objectives: To quantify behavioral complexity in probable Alzheimer’s disease (AD), compare complexity phenotypes, and determine whether behavioral complexity provides clinically meaningful information beyond total neuropsychiatric burden. We also explored whether global amyloid extent and lobar amyloid topography added explanatory value. Methods: In this cross-sectional retrospective study, we analyzed 245 psychotropic drug-naïve patients with probable AD, positive ¹⁸F-FC119S amyloid positron emission tomography (PET), and complete neuropsychiatric, cognitive, functional, and regional PET data. Behavioral complexity was derived from 12 Korean Neuropsychiatric Inventory domains using symptom count, normalized Shannon entropy of the frequency × severity profile, and a composite index. Patients were classified into tertiles. Multivariable regression and burden-stratified analyses examined associations with cognition, dementia severity, function, and amyloid measures. Results: Higher behavioral complexity was associated with lower Korean Mini-Mental State Examination (K-MMSE) scores and higher Clinical Dementia Rating (CDR) and Global Deterioration Scale (GDS) stages. In multivariable analysis, higher CDR, higher GDS, and lower Barthel Index independently predicted greater complexity, whereas amyloid extent did not. After adjustment for total neuropsychiatric burden, higher CDR remained independently associated with the composite complexity index and normalized entropy, while amyloid extent remained non-significant. Complexity-related clinical differences were most evident in the lowest burden stratum and attenuated at higher burden levels. Regional amyloid analyses yielded only selective signals. Conclusions: Behavioral complexity is a clinically meaningful neuropsychiatric phenotype in AD. Although strongly related to total neuropsychiatric burden, it is not fully reducible to it, with its clearest independent association seen for global dementia severity, particularly at lower overall burden. Full article

Early wildfire monitoring in power transmission corridors requires reliable detection of weak fire and smoke cues under complex field conditions and strict edge-computing constraints. To address these issues, this paper proposes an edge-deployable RGB–thermal framework based on visible and thermal infrared (TIR) imaging for unmanned aerial vehicle (UAV)-based corridor monitoring, including a spatial detector, YOLO-MMSC, and a temporal-enhanced version, YOLO-MMSC-T. The study also establishes a self-collected corridor-oriented RGB–thermal (RGB–T) dataset to complement public wildfire data. Unlike existing RGB–thermal wildfire datasets that mainly focus on forest or wildland fire scenes, the proposed dataset is specifically organized for complex-background power transmission-corridor monitoring, including continuous UAV sequences, nighttime conditions, smoke/vegetation occlusion, long-range small targets, and hard-negative interference. To the best of our knowledge, this is the first self-collected RGB–thermal wildfire dataset designed for this specific application scenario. The framework integrates a mobile inverted bottleneck convolution (MBConv) lightweight backbone, a Shallow Detail Fusion Module (SDFM) for shallow cross-modal alignment and denoising, a Content-Guided Attention (CGA) module for adaptive fusion, and normalized Wasserstein distance (NWD)-based box regression for long-range small-target localization. Experiments on public and self-collected datasets show that YOLO-MMSC achieves 94.6% mAP@0.5, 95.0% precision, and 93.9% recall while running at 60 FPS on Jetson Orin NX. With temporal fine-tuning, YOLO-MMSC-T reaches a continuous detection rate (CDR) of 95.6% with a jitter index of $2.8\times {10}^{-3}$. Field experiments using a DJI Matrice 4T further indicate a practical operating altitude of 120–180 m. These results support lightweight RGB–thermal remote sensing for real-time wildfire monitoring in complex transmission-corridor environments. Full article

In many developing regions, the lack of on-site weather data impedes the estimation of rainfall-driven processes, such as soil erosion. Satellite-based precipitation products (SPPs) can support hydrological modelling in gauge-sparse regions by providing continuous rainfall estimates. Accurate rainfall estimation is crucial to soil erosion modelling, particularly in data-scarce regions such as the TRB. In this study, seven satellite-based precipitation products—CHIRPS, IMERG, TRMM, ERA5, GLDAS, and PERSIANN-CDR, along with the IMD gridded dataset—were evaluated for their ability to represent rainfall patterns and support R-factor estimation in the RUSLE framework. This is the first comprehensive evaluation of multiple SPPs for RUSLE-based soil erosion modelling in the Tungabhadra river basin (TRB), providing insights for ungauged watersheds in India. CHIRPS and IMERG displayed relatively smooth and continuous patterns, while PERSIANN-CDR and TRMM exhibited fragmented rainfall zones. ERA5 and GLDAS demonstrated consistent but moderate values across the basin. IMD data served as the reference product for comparison. The findings reveal that the choice of precipitation dataset directly affects the accuracy of erosion estimation. Therefore, multi-dataset evaluation is recommended for reliable assessment of soil loss and watershed planning in ungauged or partially gauged catchments. Full article

Evaluating conditional time-series generation models remains challenging in battery research, where degradation data are often limited and experiments cover only a small number of operating conditions. The widely used Fréchet Inception Distance (FID) summarizes all conditions into a single score, which can obscure failures under rare but safety-critical conditions. Several condition-aware extensions of FID, including Conditional Fréchet Inception Distance (CFID), partially address this limitation by evaluating each condition separately. However, these approaches do not assess whether physically meaningful relationships between operating conditions are preserved, and their reliability deteriorates when only a few samples are available for each condition. To address these issues, we propose a three-layer diagnostic framework for evaluating conditional generative models under limited-data conditions. The first layer, Stratified Fréchet Distance, identifies the specific operating conditions and degradation phases where generation quality degrades. The second layer, based on Conditional Response Consistency (CRC), Conditional Distance Ratio (CDR), and Mean-Order Preservation (MOP), evaluates whether the model preserves the distance structure and ordering between conditions. MOP detects condition-ordering defects that CRC cannot identify when the real data distance matrix is non-monotone. This layer also enables statistically meaningful comparisons even when only a small number of samples are available. The third layer detects strata where statistical estimates are unreliable and provides a more stable alternative for evaluation. We validate the framework on four battery degradation datasets using two generative model architectures. The proposed approach reveals condition-specific failures that are not captured by conventional FID. It localizes generation errors to the late-stage high-temperature degradation regime that is most relevant to battery safety. The framework also detects structural distortions with statistical significance. In addition, it consistently ranks physics-informed model variants across quality differences spanning seven orders of magnitude. These results demonstrate that the proposed framework provides a practical and physically interpretable evaluation methodology for conditional generative modeling in battery degradation analysis. Full article