Search

Federated learning offers new ideas for solving the problem of multiparty joint modeling in “data silos”.Federated support vector machines can realize cross-device support vector machine modeling without local data,but the existing research has some defects such as insufficient privacy protection in a training process and a lack of research on nonlinear federated support vector machines.To solve the above problems,this paper utilizes the stochastic Fourier feature method and CKKS homomorphic encryption system to propose a nonlinear federated support vector machine training(PPNLFedSVM) algorithm for privacy protection.Firstly,the same Gaussian kernel approximate mapping function is generated locally for each participant based on the random Fourier feature method,and the training data of each participant is explicitly mapped from the low-dimensional space to the high-dimensional space.Secondly,the model parameter security aggregation algorithm based on CKKS cryptography ensures the privacy of model parameters and their contributions during the model aggregation process.Moreover,the parameter aggregation process is optimized and adjusted according to the characteristics of CKKS cryptography to improve the efficiency of the security aggregation algorithm.Security analysis and experimental results show that the PPNLFedSVM algorithm can ensure the privacy of participant model parameters and their contributions to the training process without losing the model accuracy.

A lightweight water-soluble core mold material for hollow shell composites was prepared by hot drying process with polyvinylpyrrolidone (PVP) aqueous solution as the adhesive and hollow glass microspheres as the matrix material. The effects of the content of adhesive and different treatment methods on the flexural strength, compressive strength and dissolution rate of the core mold were studied, and the microstructure of the core mold material was observed by scanning electron microscope. The results show that adhesive dosage of 40%, the flexural strength and compressive strength could reach up to 3.346 MPa and 3.499 MPa, with a water solubilization rate of 0.238 g/s. After the coupling agent treatment and high temperature treatment, the core mold strength and dissolution rate are improved, and the modification effect is good; small filler sizes can effectively fill the gap formed by matrix materials accumulation, improve the mechanical strength and affect the water solubility rate. The properties of water-soluble core can meet the requirements of resin transfer molding (RTM) and vacuum bag molding, which provides a solution for the integrated molding of profiled hollow profiles. [ABSTRACT FROM AUTHOR]

A lightweight and high-strength water-soluble core mold for composites was prepared by hot-air drying method using aqueous solution of polyvinylpyrrolidone as the adhesive and hollow glass beads, gypsum and graphite as the matrix materials. The effects of adhesive content and matrix material on the mechanical properties and dissolution rates of the water-soluble core mold were investigated. The effect of modification methods on the mechanical properties of water-soluble core molds was investigated, and the microscopic morphology of the modified water-soluble core mold was observed by scanning electron microscopy (SEM), and its modification mechanism was initially explored. The results shows that the density of water-soluble core mold is 0.46~0.58 g/cm³, and the high-temperature treatment and the compounding of glass beads with different scales of particle size are the effective means of modification. The flexural strength of the water-soluble core can reach 7.46 MPa, the compression strength reaches 7.45 MPa, and the dissolution rate is 1.11 g/min before the modification, while the flexural strength of the modified core mold can reach 13.21 MPa, the compression strength reaches 13.45 MPa, and the dissolution rate is 6.34 g/min. The comprehensive performance of the water-soluble core mold can meet the requirements of the composite molding process, and it provides a solution for the integrated molding of the hollow profiled composites. [ABSTRACT FROM AUTHOR]

This study explores whether guanidinoacetic acid (GAA) addition can regulate nutrient degradability, rumen fermentation characteristics, and gas composition in two sheep-fattening diets. A 2 × 8 factorial in vitro culture was examined to determine the effects of GAA addition at the following levels of 0%, 0.03%, 0.05%, 0.07%, 0.09%, 0.11%, 0.13%, and 0.15% of two total mixed rations (T1 diet: early fattening stage diet; T2 diet: late fattening stage diet). After 72 h in vitro incubation of two diets with mixed rumen liquid obtained from six rumen-cannulated lambs, the T2 diet exhibited higher dry matter (DM) digestibility, higher cumulative gas production at 72 h (GP72), higher asymptotic gas production(A), and longer the time at which half of A is reached (C). However, it exhibited a lower acetic acid and a lower ratio of acetate to propionate than the diet of T1. A quadratic increase occurred in neutral detergent fiber (NDF) and acid detergent fiber (ADF) digestibility, with a maximum point occurring at the 0.09% GAA group. The gas production kinetic result indicated that increasing the level of GAA addition resulted mainly in an increase of GP72 and A, with the maximum point occurring at 0.09% for the T1 diet and 0.07–0.09% for the T2 diet. Moreover, the levels of GAA addition did not affect pH, the proportion of any of the volatile acid, or gas composition, but when the levels of GAA addition were increased, the microbial crude protein (MCP), ammonia nitrogen (NH3-N), and total volatile fatty acid (TVFA) content exhibited a quadratic relationship. The highest MCP contents were seen in the 0.07%, 0.09%, and 0.11% groups, while NH3-N and TVFA were in the 0.07% group. In summary, the appropriate level of GAA addition in early and late fattening stage diets ranged from 0.07% to 0.11%. [ABSTRACT FROM AUTHOR]

Abstract Background Although socioeconomic status (SES) has been focused on as a key determinant of cancer stage at diagosis in western countries, there has been no systemic study on the relationship of SES and breast cancer stage at diagnosis in China. Methods The medical charts of 4,211 eligible breast cancer patients from 7 areas across China who were diagnosed between 1999 and 2008 were reviewed. Four area-based socioeconomic indicators were used to calculate area-based SES by cluster analysis. The associations between area-based SES and stage at diagnosis were analyzed by trend chi-square tests. Binary logistic regression was performed to estimate odds ratios for individual demographic characteristics' effects on cancer stages, stratified by area-based SES. Results The individual demographic and pathologic characteristics of breast cancer cases were significantly different among the seven areas studied. More breast cancer cases in low SES areas (25.5%) were diagnosed later (stages III & IV) than those in high (20.4%) or highest (14.8%) SES areas (χ2 for trend = 80.79, P < 0.001). When area-based SES is controlled for, in high SES areas, cases with less education were more likely to be diagnosed at later stages compared with more educated cases. In low SES areas, working women appeared to be diagnosed at earlier breast cancer stages than were homemakers (OR: 0.18-0.26). Conclusions In China, women in low SES areas are more likely to be diagnosed at later breast cancer stages than those in high SES areas.

Abstract Background According to the very limited cancer registry, incidence and mortality rates for female breast cancer in China are regarded to be increasing especially in the metropolitan areas. Representative data on the breast cancer profile of Chinese women and its time trend over years are relatively rare. The aims of the current study are to illustrate the breast cancer profile of Chinese women in time span and to explore the current treatment approaches to female breast cancer. Methods This was a hospital-based nation-wide and multi-center retrospective study of female primary breast cancer cases. China was divided into 7 regions according to the geographic distribution; from each region, one tertiary hospital was selected. With the exception of January and February, one month was randomly selected to represent each year from year 1999 to 2008 at every hospital. All inpatient cases within the selected month were reviewed and related information was collected based on the designed case report form (CRF). The Cancer Hospital/Institute, Chinese Academy of Medical Sciences (CICAMS) was the leading hospital in this study. Results Four-thousand two-hundred and eleven cases were randomly selected from the total pool of 45,200 patients and were included in the analysis. The mean age at diagnosis was 48.7 years (s.d. = 10.5 yrs) and breast cancer peaked in age group 40-49 yrs (38.6%). The most common subtype was infiltrating ductal carcinoma (86.5%). Clinical stage I & II accounted for 60.6% of 4,211 patients. Three-thousand five-hundred and thirty-four cases had estrogen receptor (ER) and progestin receptor (PR) tests, among them, 47.9% were positive for both. Two-thousand eight-hundred and forty-nine cases had human epidermal growth factor receptor 2(HER-2) tests, 25.8% of them were HER-2 positive. Among all treatment options, surgery (96.9% (4,078/4,211)) was predominant, followed by chemotherapy (81.4% (3,428/4,211). Much less patients underwent radiotherapy (22.6% (952/4,211)) and endocrine therapy (38.0% (1,599/4,211)). Conclusions The younger age of breast cancer onset among Chinese women and more advanced tumor stages pose a great challenge. Adjuvant therapy, especially radiotherapy and endocrine therapy are of great unmet needs.