Protein–Protein Interactions Prediction Base on Multiple Information Fusion via Graph Representation Learning

It is a critical impact on the processing of biological cells to protein–protein interactions (PPIs) in nature. Traditional PPIs predictive biological experiments consume a lot of human and material costs and time. Therefore, there is a great need to use computational methods to forecast PPIs. Most of the existing calculation methods are based on the sequence characteristics or internal structural characteristics of proteins, and most of them have the singleness of features. Therefore, we propose a novel method to predict PPIs base on multiple information fusion through graph representation learning. Specifically, firstly, the known protein sequences are calculated, and the properties of each protein are obtained by k-mer. Then, the known protein relationship pairs were constructed into an adjacency graph, and the graph representation learning method–graph convolution network was used to fuse the attributes of each protein with the graph structure information to obtain the features containing a variety of information. Finally, we put the multi-information features into the random forest classifier species for prediction and classification. Experimental results indicate that our method has high accuracy and AUC of 78.83% and 86.10%, respectively. In conclusion, our method has an excellent application prospect for predicting unknown PPIs.

Multi-Objective Optimization Information Fusion and Its Applications for Logistics Centers Maximum Coverage

From past the development direction of logistics centers covering problem, the main solution is almost always relying on modern computer and gradually developed intelligent algorithm, at the same time, the previous understanding of dynamic covering location model is not "dynamic", in order to improve the unreasonable distribution of logistics centers deployment time, improve the service coverage, coverage as the optimization goal to logistics centers, logistics centers as well as each one can be free to move according to certain rules of "dot", according to the conditions set by the site moved to a more reasonable. The innovation of all algorithms in this paper lies in that the logistics centers themselves are regarded as the subject of free "activities", and they are allowed to move freely according to these rules by setting certain moving rules. Simulation results show that the algorithm has good coverage effect and can meet the requirements of logistics centers for coverage effect.

Collaborative Reflection-Augmented Autoencoder Network for Recommender Systems

As the deep learning techniques have expanded to real-world recommendation tasks, many deep neural network based Collaborative Filtering (CF) models have been developed to project user-item interactions into latent feature space, based on various neural architectures, such as multi-layer perceptron, autoencoder, and graph neural networks. However, the majority of existing collaborative filtering systems are not well designed to handle missing data. Particularly, in order to inject the negative signals in the training phase, these solutions largely rely on negative sampling from unobserved user-item interactions and simply treating them as negative instances, which brings the recommendation performance degradation. To address the issues, we develop a C ollaborative R eflection-Augmented A utoencoder N etwork (CRANet), that is capable of exploring transferable knowledge from observed and unobserved user-item interactions. The network architecture of CRANet is formed of an integrative structure with a reflective receptor network and an information fusion autoencoder module, which endows our recommendation framework with the ability of encoding implicit user’s pairwise preference on both interacted and non-interacted items. Additionally, a parametric regularization-based tied-weight scheme is designed to perform robust joint training of the two-stage CRANetmodel. We finally experimentally validate CRANeton four diverse benchmark datasets corresponding to two recommendation tasks, to show that debiasing the negative signals of user-item interactions improves the performance as compared to various state-of-the-art recommendation techniques. Our source code is available at https://github.com/akaxlh/CRANet.

Enterprise Financial Data Sharing Based on Information Fusion Cloud Computing Environment

At present, many companies have many problems such as high financial costs, low financial management capabilities, and redundant frameworks; at the same time, the SASAC requires that the enterprise’s financial strategy transfer from “profit-driven” to “value-driven”, finance separate from accounting to improve the operational efficiency of the company. Under this background, more and more enterprise respond to the call of the SASAC; in order to achieve the goals of corporate financial cost savings and financial management efficiency improved, we began to provide services through financial sharing. The research of information fusion theory involves many basic theories, which can be roughly divided into two large categories from the algorithmic point of view: probabilistic statistical method and artificial intelligence method. The main task of artificial intelligence is to realize the computer for some learning, thinking process, and wisdom formation of simulation, and an important goal of information integration is the human brain comprehensive processing ability simulation, so artificial intelligence method will have broad application prospects in the field of information fusion; the common methods have D-S evidence reasoning, fuzzy theory, neural network, genetic algorithm, rough set, and other information fusion methods. The purpose of this paper is to proceed from the internal financial situation of the enterprise, analyze data security issues in the operation of financial shared services, and find a breakthrough in solving problems. But, with constantly expanding of enterprise group financial sharing service scale, the urgent problem to be solved is how to ensure the financial sharing services provided by enterprises in the cloud computing environment. This paper combines financial sharing service theory and information security theory and provides reference for building financial sharing information security for similar enterprises. For some enterprise that have not established a financial shared service center yet, they can learn from the establishment of the financial sharing information security system in this paper and provide a reference for enterprise to avoid the same types of risks and problems. For enterprise that has established and has begun to practice a financial shared information security system, appropriate risk aversion measures combined with actual situation of the enterprise with four dimensions related to information security system optimization was formulated and described in this paper. In summary, in the background of cloud computing, financial sharing services have highly simplified operational applications, and data storage capabilities and computational analysis capabilities have been improved greatly. Not only can it improve the quality of accounting information but also provide technical support for the financial sharing service center of the enterprise group, perform financial functions better, and enhance decision support and strategic driving force, with dual practical significance and theoretical significance.