Deep neural networks trained under severe class imbalance often exhibit degraded performance, typically attributed to statistical bias. In this work, we identify a complementary optimization-level pathology: inter-class gradient interference within shared representations, where gradients from majority classes suppress minority-class learning. To analyze this phenomenon, we introduce a diagnostic framework based on layer-wise gradient flow analysis and a Gradient Conflict Matrix, which quantifies interference using cosine similarity between class-specific gradients. Using this framework, we study multi-branch convolutional architectures and propose a lightweight modification, Class-Specific Branch Attention (CSBA), that enables branch-specific channel reweighting to reduce gradient coupling. This mechanism promotes implicit feature decoupling across branches while preserving architectural simplicity. Empirically, CSBA improves minority-class performance, increasing the F1 score for the Physical-Damage class from 0.261 to 0.522 under severe imbalance, while maintaining comparable overall accuracy. Validation on CIFAR-10-LT confirms that this behavior generalizes across imbalanced visual recognition settings, with Macro-F1 improving from 0.595 to 0.655. More broadly, our findings highlight the importance of considering optimization dynamics alongside statistical methods when designing architectures for imbalanced learning.
This study presents the first explainable machine learning (ML) framework for predicting drying shrinkage in treated recycled aggregate concrete, addressing a critical knowledge gap hindering structural applications. A comprehensive database of 460 concrete mixes from 30 peer-reviewed studies (2015-2024) was compiled, encompassing seven recycled concrete aggregates (RCA) treatment strategies and replacement levels of 25%-100%. Five ML algorithms were optimized and integrated with explainable Absorption Index techniques (SHapley Additive Explanations, Local Interpretable Model-Agnostic Explanations, partial dependence plots) to provide both predictive accuracy and mechanistic transparency. Extreme Gradient Boosting achieved exceptional performance (R-2 = 0.974, root mean square error = 12.25 mu epsilon, mean absolute error = 9.29 mu epsilon), reducing prediction errors by 55% compared to classical formulations (ACI 209R, B3, Comite Euro-International du Beton - Federation Internationale de la Precontrainte [CEB-FIP]) while maintaining errors below experimental measurement uncertainty (+/- 20-30 mu epsilon). Explainability analyses revealed that curing age and natural aggregate content dominate shrinkage evolution (>85% variance), while treatment-specific thresholds were identified: mechanical and thermal treatments remain effective up to similar to 400 kg/m(3) (similar to 50% replacement), whereas chemical treatments should be limited to <700 kg/m(3). A graphical user interface enables real-time scenario analysis for performance-based mix optimization. Unlike traditional empirical codes, this framework provides instance-level explanations of how specific parameters influence shrinkage, transforming treated recycled aggregates from waste substitutes into engineered materials. The system achieves shrinkage prediction within +/- 5 mu epsilon of experimental observations for 50% RCA replacement mixtures, demonstrating that treated recycled aggregates can match conventional concrete performance while reducing embodied CO2 by 12%-18% and diverting 400-600 million tons of construction waste annually if adopted globally. The framework is validated against international design standards (ACI 209R, Eurocode 2, IS 1343), enabling immediate implementation in structural engineering practice.
The rising global burden of type 2 diabetes mellitus (T2DM) highlights the need for affordable and sustainable dietary strategies for its prevention and management. Concurrently, fruit processing industries generate substantial quantities of nutrient–rich agri–food byproducts, including peels, seeds, pomace, and pulp, which are often underutilized despite their high bioactive compounds. This review examines the potential of fruit–derived by–products as sources of antidiabetic bioactives, focusing on their compositional characteristics, key biological mechanisms, and prospects for incorporation into functional foods. Bioactive compounds such as ellagitannins, gallic acid, mangiferin, and naringin are discussed in relation to their ability to modulate carbohydrate metabolism, insulin signaling pathways, oxidative stress, and gut microbiota. Emphasis is placed on high–impact by–products, including pomegranate peel, grape pomace, apple pomace, and citrus albedo. Evidence from in vitro, animal, and limited human studies is critically evaluated, recognizing that clinical validation in humans remains emerging. The review further addresses challenges related to bioavailability, food formulation, safety, and regulatory compliance, noting that most functional food applications are currently confined to laboratory– or pilot–scale development. Finally, the role of fruit by–product valorization within circular economy (CE) frameworks is highlighted, and future research directions are outlined, including structure–function relationships, well–designed clinical trials, and the development of standardized formulations to support targeted dietary interventions for T2DM.
Sensitive and accurate detection of biologically significant bilirubin is of great importance in the field of clinical diagnostics and disease monitoring. The present study introduces the TX-100 based nanomicellar fluorescent sensing platform 3@TX-100 comprising of 1,8-naphthalimide possessing schiff base and pyrene moiety for the selective and ultrasensitive detection of bilirubin in aqueous solutions. The proposed sensor 3@TX-100 exhibited an excellent “turn-off” fluorescence response toward bilirubin with an ultralow detection limit (LOD) of 0.68 nM. A concentration-dependent decrease in the emission intensity of 3@TX-100 was noted upon the addition of bilirubin, with saturation achieved at 75 µM of bilirubin. Notably, the system displayed approximately 29-fold fluorescence quenching (≈ 96.6
High-frequency regeneration through shoot organogenesis and/or somatic embryogenesis is a prerequisite for genetic transformation of plants. Cytokinins have been mainly implicated in shoot organogenesis. The present work is focused on the effect of auxins alone on shoot organogenesis in two elite clones (‘CE2’ and ‘Y8’) of Eucalyptus tereticornis, cultivars of Solanum tuberosum, Chlorophytum borivilianum, and Gladiolus hybridus. Leaf segments taken from microshoots of these plants showed differential shoot organogenesis on medium supplemented with only auxins except in case of S. tuberosum where leaves were unable to regenerate. In S. tuberosum shoot organogenesis was induced on medium supplemented with auxins only from internodal explants. Out of the three different auxins tried (NAA, IAA, and 2, 4-D), the highest frequency of shoot organogenesis (66.67
