Graphic Era Hill University

Mango (Mangifera indica) is among the most commercially important of the tropical fruits and is very susceptible to numerous foliar infections that affect production, quality of mango fruit and also the lives of scanty farmers. It is important that these diseases be detected early and accurately so as to prevent crop disease and in a way that makes agricultural practices sustainable. In this research, a fine-tuned deep learning model- ResNet50 was created and tested on a Kaggle dataset which has more than 4000 mango leaves images in seven categories: Anthracnose, Bacterial Canker, Cutting Weevil, Die Back, Gall Midge, Powdery Mildew and Healthy leaves. Preprocessing, augmentation and priori division into training, validation, and testing subsets were done to increase the robustness of the models. The experimental findings proved that the overall accuracy was 91% with Powdery Mildew giving the highest F1-score of 0.951 and the remaining classes have balanced performances. The confusion matrix also supported credible classification and had minimal errors of misclassification among the visually similar diseases. The convergence and generalization of the model was confirmed with training and validation curves, and a stable loss decrease and steady improvement in accuracy through 80 epochs. The given system is beneficial as it generates computer-vision research in precision agriculture and also promotes the sustainability by minimizing excessive pesticide usage, enhancing food security, and increasing agricultural adaptability to climate change. This study highlights the possibility of AI-based research to benefit a healthier production and secure long-term agricultural output.

The Signaling Lymphocytic Activation Molecule (SLAM) family receptors play essential roles in regulating immune cell activation, differentiation, and communication. SLAMF5, also known as CD84, has drawn increasing attention in cancer immunology due to its involvement in both tumor progression and immune modulation. This review explores the expression patterns, signaling mechanisms, and functional roles of SLAMF5/CD84 within the tumor microenvironment. SLAMF5/CD84 is expressed on multiple immune cell types and contributes to immune evasion by enhancing regulatory B cell function, promoting myeloid-derived suppressor cell expansion, and upregulating immune checkpoint molecules such as PD-L1. Its expression has been implicated in various hematologic malignancies and solid tumors, including chronic lymphocytic leukemia, multiple myeloma, and triple-negative breast cancer. Emerging therapeutic approaches targeting SLAMF5/CD84-such as monoclonal antibodies and CAR T-cell therapies-offer promising strategies to counteract immunosuppression and improve treatment outcomes. By highlighting recent findings and therapeutic developments, this review underscores the significance of SLAMF5/CD84 as both a prognostic biomarker and a novel target in cancer immunotherapy. Understanding SLAMF5/CD84's multifaceted roles in the tumor immune landscape could support the development of more effective and personalized cancer treatment strategies.

This paper describes a system for recommending music based on real-time analysis of user facial expressions and emotional states. Facial features are recognized and tracked using a Haar Cascade Classifier, while a convolutional neural network (CNN) trained on the FER-2013 dataset classifies emotions with an 80% accuracy rate for the dominant emotional states of Happy and Surprise. There are still difficulties with the detection of more subtle emotional states and the system misclassifies and overfits with the states of Sad and Neutral. Data augmentation alleviates the problem somewhat, but the results are still suboptimal. After the emotional state is classified, music corresponding to the user’s emotional state is queued and played (i.e. uplifting music is played for Happy and soothing Sad). This voice feature of the system, however, contains potential that has yet to be harnessed. For a more personalized and interactive experience, the system could also incorporate music articulation, improvisation, and generative algorithms. Future work will include multi-modal emotion recognition, network depth for emotion analysis, and integration with dynamic libraries such as Spotify for more adaptable and versatile emotion-based music recommendation systems.

The ability to scale within hybrid and multi cloud infrastructures enhances liquid resource availability, However, this also dramatically increases the attack surface, leaving trivially configured defenses completely obsolete, particularly for attack sophistication levels identified as low to mid. Deploying Large Language Models brings unique risks, such as prompt injection and exposing unreasoned system vulnerability gaps. This is the focus of the research: integrating LLM-Guarded Clouds with preemptive, nonlinear, adaptive threat defense hybrid cloud frameworks with strong attack surface generative AI. This LLM system manages multi-stage attacks with GNNs, feeds LLMs through Controlled Authoring Tier Restricted Processor/Graph-Reduced PAXs for reasoning Focused Inferencing, and applies federated learning within TEEs for classified, low-surveillance, and safe reasoning operationally secure LLM inferencing. Our analysis of the UNSW-NB15, CIC-IDS2018 and AWS CloudTrail logs demonstrates an accuracy rate 96.5%, F1 score of 95.4% and 3.4% false positive rate – outperforming ML, CNN-LSTM and LLM only baselines by almost 15%. Robustness testing indicates 72% improvement in adversarial resilience against prompt injection. Overall, LLM-Guarded Clouds improves accuracy, resilience, and explainability in detection, providing a novel automated defense solution for hybrid cloud environments.

Abstract Earthquake Early Warning (EEW) systems are important for mitigating casualties and damage to infrastructure by providing seconds of advance notice of damaging seismic waves. However, established EEW systems depend on dense seismic networks, handcrafted features, or multi-station triangulation to find an earthquake’s P-wave arrival time requiring substantial financial investment, time, and infrastructure unsuitable for resource-constrained areas. The contribution of this paper is to propose a hybrid single-station framework, which applies physics-informed preprocessing, deep learning, and probabilistic modelling. The tri-axial accelerometer signals are pre-processed via double integration and bandpass filtering and analysed via a U-Net + + encoder–decoder with dilated convolutions and Multi-Head Self-Attention (MHSA). This allows the U-Net + + architecture to simultaneously fine-tune recognition of spatial-temporal features and to model the global dependency context necessary for accurate P-wave detection. Gaussian label smoothing has been incorporated to adapt and enhance robustness to potential uncertainty in the annotation labels, while a Bayesian Markov Chain Monte Carlo (MCMC) method provides an easy procedure for obtaining probabilistic (and noise-invariant) magnitude estimation. The evaluation across two datasets: the Stanford Earthquake Dataset (STEAD), and IoT-based simulation results illustrate significant margins of improvement, e.g., approximately 10–12% greater F1-scores for P-wave detection, and an approximately 35% lower rate of magnitude-estimation errors. By demonstrating noteworthy improvements in earthquake detection and magnitude estimation, the hybrid single-station framework is a cost-effective and real-time EEW with multiple pathways for deployment at a global scale.

Harnessing microorganisms and their metabolites to prevent diseases, offers an alternative in disease management without facing the drawbacks of chemical control. Aquatic hyphomycetes, members of ‘Fungi Imperfecti’ are the prolific producers of secondary metabolites useful in medicinal, industrial, and agricultural areas. They are the major microbial element in aquatic ecosystem occurring as active colonizer of submerged decaying leaf litter. Bioactivity of two potential aquatic hyphomycetes was tested against three phytopathogenic fungi using ‘Dual Culture’ and ‘Agar Well Diffusion’ method. Aquatic hyphomycetes were identified on the basis of morphological characters through pertinent literature while the pathogenic fungal cultures were outsourced to Bio edge solutions, Bangalore for molecular using 18s rRNA sequencing for their species confirmation. C. aquaticum showed maximum percent inhibition against F. oxysporum (49.3%) while minimum percent inhibition against F. solani (14%) in case of dual culture method. In agar well diffusion method C. aquaticum showed maximum percent inhibition against F. oxysporum (46.8%) while minimum percent inhibition against R. solani (14%). B. rhombica showed maximum percent inhibition against F. oxysporum (34%) while minimum percent inhibition against R. solani (13.5%) in dual culture method. In agar well diffusion method B. rhombica showed maximum percent inhibition against F. solani (35.4%) while minimum percent inhibition against F. oxysporum (21%). The MIC of C. aquaticum was recorded to be 50 µg/ml against F. oxysporum and 25 µg/ml against F. solani and R. solani while MIC of B. rhombica was recorded to be 50 µg/ml against all the test fungi. Minimum Inhibitory Concentration and Activity index of hyphomycetous extracts indicated their potential more or less equivalent to that of Carbendazim, a commercial fungicide.

This study focuses on the synthesis of a novel magnetic interphase palladium catalyst immobilized on pyromellitic dianhydride (PM)-coated magnetic SnFe2O4 nanoparticles. Such surface functionalization of magnetic particles represents a promising strategy to bridge the gap between heterogeneous and homogeneous catalysis methods. The structure, morphology, and physicochemical properties of these particles were thoroughly examined using various analytical techniques, including FT-IR, SEM, XRD, VSM, ICP, and EDS. The resulting SnFe2O4/SiO2/PM-Pd nanocatalyst exhibited excellent catalytic performance as a recyclable catalyst in Suzuki-Miyaura cross-coupling reactions at room temperature. Additionally, the catalyst demonstrated high reusability, showing minimal palladium leaching and no significant loss in activity across multiple cycles.Supplementary InformationThe online version contains supplementary material available at 10.1038/s41598-025-16233-9.

Emerging evidence reveals that interactions between the nervous system and tumor biology significantly influence cancer progression, metastasis, and therapeutic outcomes. Here, we elucidate the neurobiological mechanisms that underpin tumor development, highlighting the dynamic role of neural components within the tumor microenvironment (TME). Neural signals and structural adaptations in the TME stimulate tumorigenesis and enable cancer cell plasticity, mimicking neurodevelopmental processes. Astrocytic glial cells release neurotrophic factors that support metastatic colonization and enhance tumor cell survival. Notably, cancer cells can establish pseudo-tripartite synapses with neurons, promoting both proliferation and invasion. We explore the cancer-neural network interplay, emphasizing how axonal remodeling, circuit reorganization, and synaptic dysfunction not only drive tumor growth but also contribute to associated symptoms like seizures and chronic pain. Molecularly, mutations such as in PIK3CA and abnormalities in neurotransmitter signaling reveal how neurotumors communicate and adapt. Furthermore, metabolic stress responses from tumor cells can activate nociceptive neurons, sustaining malignant progression. Understanding these neurobiological interactions opens avenues for novel therapeutic strategies. Precision neuro-oncology may benefit from targeting neurotrophic signaling, synaptic pathways, and neuronal differentiation programs. Advances in biomarker research from neurotumors also contribute to improved diagnostic and prognostic tools. By integrating neuroscience insights into oncological frameworks, we propose a paradigm shift toward therapies that intercept the neural circuitry sustaining malignancies. This neuro-oncological approach holds promise in addressing aggressive cancer phenotypes and improving patient outcomes.

Liquiritin (LIQ), a bioactive flavonoid from Glycyrrhiza species, has shown significant potential in cancer therapy. LIQ exhibits potent inhibitory effects on various cancer cell types, including breast, lung, liver, and colon cancers, while demonstrating low toxicity towards healthy cells. Its anticancer mechanisms include inducing cell cycle arrest, promoting apoptosis, and modulating inflammation-related pathways. Additionally, LIQ impedes angiogenesis and enhances the efficacy of conventional chemotherapies through sensitization and synergistic effects with other natural compounds and targeted therapies. These multifaceted actions highlight LIQ as a promising candidate for further development as an anticancer agent. This abstract provides an overview of LIQ's chemistry, biological effects, and underlying mechanisms.