A Multimodal Network Security Framework for Healthcare Based on Deep Learning.

As the network is closely related to people's daily life, network security has become an important factor affecting the physical and mental health of human beings. Network flow classification is the foundation of network security. It is the basis for providing various network services such as network security maintenance, network monitoring, and network quality of service (QoS). Therefore, this field has always been a hot spot of academic and industrial research. Existing studies have shown that through appropriate data preprocessing techniques, machine learning methods can be used to classify network flows, most of which, however, are based on manually and expert-originated feature sets; it is a time-consuming and laborious work. Moreover, only features extracted by a single model can be used in classification tasks, which can easily make the model inefficient and prone to overfitting. In order to solve the abovementioned problems, this study proposes a multimodal automatic analysis framework based on spatial and sequential features. The framework is completely based on the deep learning method and realizes automatic extraction of two types of features, which is very suitable for processing large-flow information; this improves the efficiency of network flow classification. There are two types of frameworks based on pretraining and joint-training, respectively, with analyzing the advantages and disadvantages of them in practice. In terms of evaluation, compared with the previous methods, the experimental results show that the framework has good performance in both accuracy and stability.

[Regional environment, family's endowment and straw disposal behavior: A case study of dry farming areas in Gansu Province, China]. / 区域环境、家庭禀赋与秸秆处置行为­­ä»¥ç”˜è‚ƒçœæ—±ä½œå†œä¸šåŒºä¸ºä¾‹.

Understanding the treatment and influencing factors of straw is important to improve the utilization efficiency of straw resources and alleviate the negative external effects of the environment. Here, we proposed an analysis framework of farmers' straw disposal behavior based on ecological rationality. The Logit model was used to analyze the farmers' willingness and influencing factors for the selection of straw burning and feed utilization with a dataset of 424 valid questionnaires in dry farming areas of Gansu Province. The results showed that the straw disposal behavior of farmers was the result of decision-making cognition formed in the process of long-term adaptation and co-evolution between farmers and the surrounding environment. In dry farming area, the straw treatment methods were diversified, with feed utilization and fuel as the two main forms and straw incineration and discard being ubiquitous. Among the factors that affect farmers' straw burning behavior, householder age (P＜0.1), education level (P＜0.01), the scale of livestock raising (P＜0.05), the proportion of agricultural income (P＜0.1), and government policy propaganda (P＜0.01) had significant inhibitory effect. The gender of householder (P＜0.1) and cognition level (P＜0.01) helped farmers to choose the non-pro-environmental behavior, and the environmental awareness of farmers was weak. In terms of straw feed utilization behavior, householder age (P＜0.1), education level (P＜0.05), the situation of family members serving as village cadres (P＜0.05), feed proces-sing technical guidance (P＜0.01) and subsidy for prohibition of straw burning (P＜0.1) had positive effect on straw forage utilization, while the topography (P＜0.1) had a negative effect. Some policy recommendations were given to promote utilization of straw resources: constructing a combination mechanism of "prohibition of burning and subsidies", strengthening the extension of straw feed utilization technology, and accelerating the improvement of straw collection-store-transportation service system.

[Effects of different tillage measures on upland soil respiration in Loess Plateau].

A field experiment was conducted in Lijiabu Town of Dingxi City, Gansu Province to study the soil respiration and its relations with the canopy temperature and soil moisture content in a rotation system with spring wheat and pea under effects of different tillage measures. Six treatments were installed, i.e., tillage with no straw- or plastic mulch (conventional tillage, T), tillage with straw mulch (TS), tillage with plastic mulch (TP), no-tillage (NT), no-tillage with straw mulch (NTS), and no-tillage with plastic mulch (NTP). During the growth periods of spring wheat and pea, soil respiration had different change patterns, with the peaks appeared at the early jointing, grain-filling, and maturing stages of spring wheat, and at the 5-leaf, silking, flowering and poding, in spring wheat field between treatments NTS and T, and the soil respiration rate was significantlyand maturing stages of pea. There was an obvious difference in the diurnal change of soil respiration lower in NTS than in T; while the soil respiration in pea field had less diurnal chan ge. Soil respiration rate had a significant linear relationship with the canopy temperature of both spring wheat andpea, the correlation coefficient being the highest at booting stage of spring wheat and at flowering and poding stage of pea, followed by at grain-filling stage of spring wheat and at branching stage of pea. There was also a significant parabola relationship between soil respiration rate and soil moisture content, the correlation coefficient being higher under conservation tillage than under conventional tillage, with the highest under NTS. The moisture content in 10-30 cm soil layer of spring wheat field and that in 5-10 cm soil layer of pea field had the greatest effects on soil respiration. Comparing with conventional tillage, all the five conservation tillage measures decreased soil respiration, with the best effects of no-tillage with straw mulch.