Output uncertainty indicates whether the probabilistic properties reflect objective characteristics of the model output. Unlike most loss functions and metrics in machine learning, uncertainty pertains to individual samples, but validating it on individual samples is unfeasible. When validated collectively, it cannot fully represent individual sample properties, posing a challenge in calibrating model confidence in a limited data set. Hence, it is crucial to consider confidence calibration characteristics. To counter the adverse effects of the gradual amplification of the classifier output amplitude in supervised learning, we introduce a post-processing parametric calibration method, $\rho$-Norm Scaling, which expands the calibrator expression and mitigates overconfidence due to excessive amplitude while preserving accuracy. Moreover, bin-level objective-based calibrator optimization often results in the loss of significant instance-level information. Therefore, we include probability distribution regularization, which incorporates specific priori information that the instance-level uncertainty distribution after calibration should resemble the distribution before calibration. Experimental results demonstrate the substantial enhancement in the post-processing calibrator for uncertainty calibration with our proposed method.
Microservice is an architecture style that decomposes complex software into loosely coupled services, which could be developed, maintained, and deployed independently. In recent years, the microservice architecture has been drawing more and more attention from both industrial and academic communities. Many companies, such as Google, Netflix, Amazon, and IBM have applied microservice architecture in their projects. Researchers have also studied microservices in different directions, such as microservices extraction, fault localization, and code quality analysis. The recent work has presented cross-service code clones are prevalent in microservice projects and have caused considerable co-modifications among different services, which undermines the independence of microservices. But there is no systematic study to reveal the underlying reasons for the emergence of such clones. In this paper, we first build a dataset consisting of 2,722 pairs of cross-service clones from 22 open-source microservice projects. Then we manually inspect the implementations of files and methods involved in cross-service clones to understand why the clones are introduced. In the file-level analysis, we categorize files into three types: DPFile (Data-processing File), DRFile (Data-related File), and DIFile (Data-irrelevant File), and have presented that DRFiles are more likely to encounter cross-service clones. For each type of files, we further classify them into specific cases. Each case describes the characteristics of involved files and why the clones happen. In the method-level analysis, we dig information from the code of involved methods. On this basis, we propose a catalog containing 4 categories with 10 subcategories of method-level implementations that result in cross-service clones. We believe our analyses have provided the fundamental knowledge of cross-service clones, which can help developers better manage and resolve such clones in microservice projects.
Intrusion detection technology is a kind of active security technology of computer network, which can detect all kinds of malicious attacks in time and respond to the network system.It is a reasonable complement of traditional security technology, such as firewall, is a new network security technology, but also a hot spot in the research of computer network security theory.This paper mainly introduces the main function, structure model, main types and defects of intrusion detection system.Active defense is one of the basic ideas of information security.This paper gives a detailed analysis of the intrusion detection system (IDS) in the application of active defense in the shortcomings and deficiencies, further elaborated the intrusion prevention system (IPS) design idea, technical characteristics and development direction, and demonstrates the important role of IPS in active defense.Research and application show that the transition from intrusion detection to intrusion protection is the inevitable choice of information security.
Facing the problem and current situation of the power grid master-slave station dispatching system architecture, this paper using the master-slave station distributed file system and master-slave station two-level business platform to divide the traditional master station business into master-slave station interacted distribution business, also issued the master-slave station distributed integration application architecture plan. Then, discussed the technology include file system directory service, the concept of master-slave business partition, gateway proxy. Proceed further modeling analysis to the slave station business such as monitoring and control, smart alarm and also give a conclusion on apply authorization. Through the apply authorization to indicate the plan could effectively solve the current problem and the application development will have a brightened prospects.
There are certain standards to measure whether a digital campus has been successfully constructed, including.• Whether there is an advanced philosophy and a feasible plan for informationized construction.
'/%3e%3cuse xlink:href='%23a' transform='rotate(23.036 .093 25.536)'/%3e%3cuse xlink:href='%23a' transform='rotate(45.87 1.273 16.18)'/%3e%3cuse xlink:href='%23a' transform='rotate(69.945 .996 12.078)'/%3e%3cuse xlink:href='%23a' transform='rotate(20.66 -19.689 31.932)'/%3e%3c/svg%3e)