Explore the vast range of titles available.

According to the World Health Organization (WHO), pneumonia kills about 2 million children under the age of 5 every year. Traditional machine learning methods can be used to diagnose chest X-rays of pneumonia in children, but there is a privacy and security issue in centralizing the data for training. Federated learning prevents data privacy leakage by sharing only the model and not the data, and it has a wide range of application in the medical field. We use federated learning method for classification, which effectively protects data security. And for the data heterogeneity phenomenon existing in the actual scenario, which will seriously affect the classification effect, we propose a method based on two-end control variables. Specifically, based on the classical federated learning FedAvg algorithm, we modify the loss function on the client side by adding a regular term or a penalty term, and add momentum after the average aggregation on the server side. The federated learning approach prevents the data privacy leakage problem compared to the traditional machine learning approach. In order to solve the problem of low classification accuracy due to data heterogeneity, our proposed method based on two-end control variables achieves an average improvement of 2% and an accuracy of 98% on average, and 99% individually, compared to the previous federated learning algorithms and the latest diffusion model-based method. The classification results and methodology of this study can be utilized by clinicians worldwide to improve the overall detection of pediatric pneumonia.

Generative AI offers new potential for creativity in marketing, yet existing systems often lack behavioral grounding and multichannel sensitivity, limiting their effectiveness in real-world business scenarios. Addressing this gap, the authors propose CAMA-GPT, a modular architecture that combines content generation with temporal user behavior modeling, lead scoring, and causal attribution analysis. The framework comprises four integrated units: a prompt-tuned creative generator, a temporal graph neural network for multichannel path modeling, a lead qualification engine, and a SHAP-based analyzer for content-value impact. Through eight comprehensive experiments using datasets such as Criteo Click Logs, a customer segmentation corpus, and a conversational recommendation dataset, CAMA-GPT demonstrates consistent improvements. It achieves an 18.2% gain in creativity score, +13.5% AUC in behavioral modeling, 4.2× top-decile lift in lead scoring, and a peak $0.63 revenue per impression in email simulations—each outperforming corresponding baselines.

Listeriosis is a zoonotic disease caused by Listeria monocytogenes (Lm), posing a significant threat to the breeding industry and public health. Ruminant livestock are particularly susceptible to Lm,  thus effective strategies are needed for controlling ovine listeriosis. In this study, we developed two inactivated vaccines and evaluated their efficacy against Lm infection in murine and ovine models. We inactivated the Lm serotype 4h XYSN strain and adjuvanted it with water-in-oil ISA 61 VG (61 VG-AIV) or aluminum (Al-AIV). Pathological observations confirmed the safety of both vaccines in mice and sheep. The immunological assays demonstrated that, compared with the Al-AIV, the 61 VG-AIV induced higher levels of Lm-specific antibodies and proinflammatory cytokines, suggesting that the ISA 61 VG adjuvant has superior immunostimulatory effects compared with the alum adjuvant. 61 VG-AIV elicited greater immunoprotection than Al-AIV (83.4% vs. 50%) against serotype 4h Lm strain challenge in mice. Additionally, 61 VG-AIV afforded cross-protection against challenges with serotypes 1/2a, 1/2b, and 4b Lm strains. Importantly, high immunoprotection in sheep was conferred by the 61 VG-AIV group (83.4%). Taken together, our findings demonstrate that the ISA 61 VG adjuvant contributes to enhancing the humoral and cellular immune responses of inactivated Lm, and 61 VG-AIV is a promising vaccine candidate for the prevention and control of animal listeriosis. This research lays a solid foundation for its application in veterinary medicine.

The co-occurrence of graphene-based nanomaterials such as reduced graphene oxide (RGO) and triclosan in agricultural soils is an emerging concern. This study investigates the impact of RGO on the formation and characteristics of bound residues (BRs) of triclosan in paddy soil using 14C-isotope tracing and LC-QTOF-MS. Results demonstrate that RGO significantly enhances the accumulation of triclosan BR in a dose-dependent manner, with the highest concentration (1.19 mg kg–1; 57.0%) observed at 500 mg kg–1 RGO. While the BR is primarily associated with the humin fraction (>63.8%), RGO shifts the distribution of 14C-triclosan, enhancing its retention in humin by 1.89–7.59% and in humic acid by 20.7–52.1%. RGO may increase the sequestered BR (8.8–24.7%), and it enhances the covalent BR of triclosan by increasing the proportions of both ether- (3.78–4.58%) and ester-bound (22.8–39.5%) forms. Metabolite analysis reveals limited transformation of triclosan (0.057–0.082 mg kg–1) in BRs, with carboxylated derivatives identified as minor products. The findings indicate that RGO enhances the persistence of triclosan BRs, which may be attributed to strong adsorption and microbial inhibition, raising concerns about their potential future remobilization and entry into the food chain. This underscores the need to assess the ecological risks of nanomaterial co-contamination for soil health and sustainable agriculture.