Many accidents happen under shunting mode when the speed of a train is below 45 km/h. In this mode, train attendants observe the railway condition ahead using the traditional manual method and tell the observation results to the driver in order to avoid danger. To address this problem, an automatic object detection system based on convolutional neural network (CNN) is proposed to detect objects ahead in shunting mode, which is called Feature Fusion Refine neural network (FR-Net). It consists of three connected modules, i.e., the depthwise-pointwise convolution, the coarse detection module, and the object detection module. Depth-wise-pointwise convolutions are used to improve the detection in real time. The coarse detection module coarsely refine the locations and sizes of prior anchors to provide better initialization for the subsequent module and also reduces search space for the classification, whereas the object detection module aims to regress accurate object locations and predict the class labels for the prior anchors. The experimental results on the railway traffic dataset show that FR-Net achieves 0.8953 mAP with 72.3 FPS performance on a machine with a GeForce GTX1080Ti with the input size of 320 × 320 pixels. The results imply that FR-Net takes a good tradeoff both on effectiveness and real time performance. The proposed method can meet the needs of practical application in shunting mode.

Recently, with the great development of e-health, more and more countries have made certain achievements in the field of electronic medical treatment. The digitization of medical equipment and the structuralization of electronic medical records are the general trends. While bringing convenience to people, the explosive growth of medical data will further promote the value of mining medical data. Obviously, finding out how to safely store such a large amount of data is a problem that urgently needs to be solved. Additionally, the particularity of medical data makes it necessarily subject to great privacy protection needs. This reinforces the importance of designing a safe solution to ensure data privacy. Many existing schemes are based on single-server architecture, which have some natural defects (such as single-point faults). Although blockchain can help solve such problems, there are still some deficiencies in privacy protection. To solve these problems, this paper designs a medical data privacy protection system, which integrates blockchain, group signature, and asymmetric encryption to realize reliable medical data sharing between medical institutions and protect the data privacy of patients. This paper proves theoretically that it meets our security and privacy requirements, and proves its practicability through system implementation.

As the core of blockchain technology, the consensus algorithm plays an important role in determining the security, data consistency, and efficiency of blockchain systems. The existing mainstream consensus algorithm is experiencing difficulties satisfying the needs of efficiency, security, and decentralization in real-world scenarios. This paper proposes a hybrid consensus algorithm based on modified Proof-of-Probability and Delegated Proof-of-Stake. In this method, the work of block generation and validation is, respectively, completed by the nodes using the modified Proof-of-Probability consensus algorithm and Delegated Proof-of-Stake consensus algorithm. When a transaction occurs, the system sends several target hash values to the whole network. Each modified Proof-of-Probability node has a different sorting algorithm, so they have different mining priorities. Every time a hash is decrypted by a modified Proof-of-Probability node, the modulo operation is done to the value of nonce, which is then compared with the expected value given by the supernode selected by the Delegated Proof-of-Stake nodes. If they are not the same, the Proof-of-Probability node enters the waiting time and the other Proof-of-Probability nodes continue to mine. By adopting two consensus algorithms, the malicious nodes must control more than 51% of the nodes that adopt the two consensus algorithms, at the same time, to effectively attack the system, that is, they must have more than 51% of the computing power and more than 51% of the tokens. This not only increases the cost of malicious attacks, but also reduces waste of computing power. In addition, the efficiency of the DPoS algorithm makes up for the deficiency of the PoP algorithm in system efficiency, and the mining behavior based on probability in the PoP algorithm also significantly weakens the ability of supernodes in the DPoS algorithm to conduct monopoly behavior or other malicious behaviors. In a word, the combination of the two algorithms makes the system perform better in terms of security, system efficiency, and decentralization.

With the rapid increase in the popularity of big data and internet technology, sequential recommendation has become an important method to help people find items they are potentially interested in. Traditional recommendation methods use only recurrent neural networks (RNNs) to process sequential data. Although effective, the results may be unable to capture both the semantic-based preference and the complex transitions between items adequately. In this paper, we model separated session sequences into session graphs and capture complex transitions using graph neural networks (GNNs). We further link items in interaction sequences with existing external knowledge base (KB) entities and integrate the GNN-based recommender with key-value memory networks (KV-MNs) to incorporate KB knowledge. Specifically, we set a key matrix to many relation embeddings that learned from KB, corresponding to many entity attributes, and set up a set of value matrices storing the semantic-based preferences of different users for the corresponding attribute. By using a hybrid of a GNN and KV-MN, each session is represented as the combination of the current interest (i.e., sequential preference) and the global preference (i.e., semantic-based preference) of that session. Extensive experiments on three public real-world datasets show that our method performs better than baseline algorithms consistently.

Scheduling EV user’s charging behavior based on charging price and applying renewable energy resources are the effective methods to release the load pressure of power grids brought about by the large-scale popularity of electric vehicles (EVs). This paper presents a novel approach for EV charging scheduling based on price negotiation. Firstly, the EV charging system framework based on price negotiation and renewable energy resources is discussed. Secondly, the price negotiation model is presented, including the initial price models and the conditions of transactions. Finally, an EV charging scheduling mechanism based on price negotiation (CSM-PN), including the price adjustment strategies of both the operator and EV users is proposed to seek a final transaction during multi-round price negotiation. Simulation results show that this novel approach can effectively improve the charging station operator’s income, reduce the EV users’ costs, and balance the load of the power grid while improving the efficiency of the EV charging system.

Session-based recommendation, which aims to match user needs with rich resources based on anonymous sessions, nowadays plays a critical role in various online platforms (e.g., media streaming sites, search and e-commerce). Existing recommendation algorithms usually model a session as a sequence or a session graph to model transitions between items. Despite their effectiveness, we would argue that the performance of these methods is still flawed: (1) Using only fixed session item embedding without considering the diversity of users’ interests and target items. (2) For user’s long-term interest, the difficulty of capturing the different priorities for different items accurately. To tackle these defects, we propose a novel model which leverages both the target attentive network and self-attention network to improve the graph-neural-network (GNN)-based recommender. In our model, we first model user’s interaction sequences as session graphs which serves as the input of the GNN, and each node vector involved in session graph can be obtained via the GNN. Next, target attentive network can activates different user interests corresponding to varied target items (i.e., the session embedding learned varies with different target items), which can reveal the relevance between users’ interests and target items. At last, after applying the self-attention mechanism, the different priorities for different items can be captured to improve the precision of the long-term session representation. By using a hybrid of long-term and short-term session representation, we can capture users’ comprehensive interests at multiple levels. Extensive experiments demonstrate the effectiveness of our algorithm on two real-world datasets for session-based recommendation.