Search

The theory of traditional Chinese medicine (TCM) is an important part of Chinese culture. In the long history, there are a large number of excellent prescriptions, whose laws have been explored by many studies, but few works directly studied the generation of prescriptions. With the rapid development of deep learning, many applications of text generation using neural networks have emerged. Prescriptions are the doctors’ clinical experience and the results of neural networks also come from the accumulated experience. So, it is very feasible to apply deep learning techniques to the recommendation on TCM prescriptions. GAN and its variants have been applied in text generation recently. It has advantages in many aspects, such as the rapid speed of computation, the update of parameters by back propagation without Markov chain and more real data generation with two-players game. We attempted to know the important attributes of prescriptions and use these contents as the training data for variants of GAN to generate prescriptions. Specifically, we attempted to apply SeqGAN (Sequence Generative Adversarial Nets) and CGAN (Conditional Generative Adversarial Nets) to prescription generations. By underlying the knowledge of TCM, the prescriptions with different characteristics can be successfully generated. In the experiments, we conducted the comparative evaluations on the original data with other models. The results showed that applications in the innovation of prescription sequence generations have certain feasibility and significance, even can provide some reference values for the innovation of TCM.

The robot kinematic error model may include redundant parameters, which require removal prior to the identification of kinematic parameters to ensure the accuracy and stability of such. Traditionally, the redundant parameters are determined by the numerical analysis. This paper presents an analytical method to determine the identifiable kinematic parameters for serial-robot calibration under various identification conditions. Among various kinematic models of robots, the DH (Denavit-Hartenberg) model is the most commonly used, so the analysis of redundant parameters is based on the DH error model. The correlation between the columns of the Jacobian matrix is analyzed by theoretical methods, resulting in the linear relationship between these columns. The redundant error parameters and the linear independent parameters after removing redundancies from robots with different measurement conditions and different configurations are clearly given. Thus, the redundant parameters can be easily determined based on the values of kinematic parameters, which facilitates the determination of the non-redundant kinematic calibration model. In addition, the physical meanings of redundant parameters are also elaborated, which can reveal the inherent principle of the error parameters’ linear correlation. Finally, the parameter identification simulation experiment shows that the identification results, after removing the redundant parameters, all converge to the true values when the random noise is not considered, which verifies the correctness of the parameter redundancy analysis results. In the case of considering random noise, the non-redundant parameter identification has better convergence than the full parameter identification. This study can provide theoretical and application references for robot kinematic calibration.

Due to the irregularity of colorectal tumor contours, it is a challenging task for clinicians to segment colorectal tumors manually in CT scans. To solve this problem, a novel algorithm based on a convolutional neural network with an attention mechanism is proposed to automatically achieve tumor segmentation. The proposed network consists of three major modules: an encoder module, which is fed CT scans to attain the feature map; a dual attention module, which includes a channel attention module and a position attention module to obtain more contextual information in the deep layer of the network; and a decoder module, which restores the feature map to the original size of the input images. We used 1131 CT slices of colorectal tumors to train and test the proposed network. Compared with U-Net and CE-Net, the Dice coefficient increased by 1.46% and 0.66% respectively, for our model. The comprehensive results show that the proposed network performs more effectively in colorectal tumor segmentation than the other methods.

The disorder, sparseness, irregularity, noise, and background of point clouds cause significant challenges in point cloud classification tasks. In such tasks, deep learning methods based on raw point cloud data have recently achieved good performance on simulated data. However, many methods experience problems when applied to realistic data containing much noise and complex background information. This paper proposes an end-to-end dual-input network (DINet) point cloud classification model based on deep learning. In the proposed model, a feature extractor obtains high-dimensional features, a feature comparator aggregates and disperses homogeneous and heterogeneous point clouds, respectively, in the feature space, and a feature analyzer completes the task. The two-channel data input facilitates a universal DINet framework that is flexible and extends to other models generalized across different datasets. DINet brings improvements in performance, presenting an overall accuracy value of 81.3% and average accuracy value of 79.6% in experiments conducted on the real-world ScanObjectNN dataset. The code for the proposed point cloud classification model is available at https://github.com/zhairf/DINet.

Under the high load, high frequency and high strength operating environment, the frequent occurrence of vehicle fault gradually attracts the attention of the society. The real-time monitoring and data recording function of vehicle-mounted equipment provides data support for vehicle status assessment and fault warning. In this paper, the real-time data collected by CAN-BUS system of Beijing Bus Group are preprocessed and discretized. On the basis of the traditional rough set theory, a new coding method is set up, and the dependency between conditional attributes and decision attributes is set as an adaptive function, which is reduced by genetic algorithm and cellular genetic algorithm respectively. The calculation results show that the key fault information of public transport vehicles is instrumental speed, oil pressure, percentage of torque, timing engine speed, and coolant temperature. By comparing the results of reduction, it is found that the cellular genetic algorithm has higher applicability than the genetic algorithm in terms of algorithm efficiency, stability, and convergence quality. Although the genetic algorithm attribute reduction is slightly better than the cellular genetic algorithm attribute reduction in the rule matching, the cellular genetic algorithm has a better ability to excavate information within the acceptable compatibility range. Finally, the selected key factors will be deployed on the Beijing Bus Group's big data platform and displayed in real time. The conclusion of this paper enriches the theory of bus engine fault warning and establishes an engine failure warning system, which can effectively reduce the failure rate of bus vehicles and reduce the maintenance cost expenditure. It has certain guiding significance for the bus operation work of Beijing Bus Group.

The subway station passenger flow prediction model can forecast passenger volume in the future. This model helps to carry out safety warnings and evacuation of passenger flow in advance. Based on the data of the Shanghai traffic card, the passenger volume in all the time intervals is clustered into three different models for prediction. Taking the Nanjing East Road Station in Shanghai as an example, the time series of passenger volumes was combined with weather data to create several supervised sequences and was converted to supervised sequences according to different values of timestep. To accelerate convergence, two artificial features were added as input. The gated recurrent unit (GRU) network model achieves accurate rolling prediction from 15 minutes to 6 hours. Finally, comparing it with the long short-term memory (LSTM) network and the back-forward propagation network (BPN), it was confirmed that the GRU network with a timestep of 1.5 hours is the best model for the long-term (more than 3 hours) traffic flow rolling prediction, while GRU with a timestep of 45 minutes has the best result for short-term rolling prediction.

To assess various segmentation methods for the anterior and posterior chambers in ultrasonic biomicroscopy (UBM) images. UBM images were collected from 102 patients diagnosed with primary angle-closure glaucoma, and the corresponding intraocular pressure (IOP) values were measured. The UBM images are manually segmented using ImageJ software as the gold standard, while automatic segmentation employs six distinct methods: Otsu threshold, K-means, fuzzy C-means, robust self-sparse fuzzy clustering algorithm, spatial intuitionistic fuzzy C-means algorithm, and fast robust fuzzy C-means (FRFCM) algorithm. The segmentation results were used to quantify the anterior chamber depth and the area, perimeter, and height of the posterior chamber. Quantitative analyses of the accuracy and reliability of the segmentation quantification results were conducted using relative error and intraclass correlation coefficients (ICC). A total of 408 clear enough UBM images were used for segmentation. The ICC values of the quantitative results of the FRFCM method are 0.996, 0.970, 0.986, and 0.970, outperforming the other five segmentation methods. It excels in accuracy, reliability, precision comparable to manual techniques, and surpasses them in reproducibility and efficiency. Thus, it can be used for semi-automatic analysis of both anterior and posterior chamber regions in UBM images. Additionally, Spearman correlation analysis was performed to explore the relationship between IOPs and the four quantification results. The Spearman correlation coefficients between IOP and anterior chamber depth, posterior chamber area, posterior chamber perimeter, and posterior chamber height were -0.0991, -0.0043, -0.0321, and 0.0757, respectively. There is no significant correlation between IOP and anterior chamber depth, posterior chamber area, perimeter, or height.

Searchable encryption is a promising technique that allows cloud service provider return the interested files to the user, even if the shared files are encrypted. However, once encrypted data is uploaded to the cloud, the data owner loses the physical control of the data; the cloud service provider may face the problem of duplicate data occupying the storage space; how can the data user obtain the legal access rights? Therefore, data integrity detection, data deduplication and reasonable access authorization become the focus of attention. To address these problems, we propose a novel verifiable attribute-based keyword search over encrypted data supporting data deduplication. In the scheme, attribute-based encryption is introduced to achieve effective access authorization based on its one-to-many advantage, as well data confidentiality. A data label of each shared file is generated to finish data deduplication, while the third-party auditor and hash function are adopted to ensure data integrity, including ciphertext, plaintext and symmetric key. Moreover, in order to solve the problem of high computational load caused by attribute-based encryption, outsourcing decryption is used to optimize the scheme. The formal security analysis shows that our scheme can implement fine-grained authorization to search and realize indistinguishable keywords, unforgeability of signatures and confidentiality of ciphertexts. Meanwhile, the results of the experiment reveal that our scheme has a low band-width communication, storage and computability.

The rapid development of communication technologies, the network, advanced computing methods and wireless medical sensors gives rise to a modern medical system. In this system, large-scale electronic health records (EHRs) are often outsourced to be stored at the third parties, such as cloud service providers (CSPs). However, CSPs are not trustworthy, that is, serious security and privacy concerns about cloud service exist because it may expose the user's sensitive data to CSPs or unauthorized users in transmission, storage and sharing. To prevent the privacy disclosure of patients better and realize information sharing more effectively, this paper proposes an anonymous EHRs sharing scheme based on decentralized hierarchical attribute-based encryption (ABE). In the proposed scheme, (1) Multiple attribute authority (AA) ABE is leveraged to achieve fine-grained and scalable data access control and avoid bottleneck. Meanwhile, hierarchical access tree is used to encrypt multiple files in one operation, thereby saving calculation and storage load greatly. Moreover, the hidden access policy enhances user privacy protection. (2) The global identifier (GID) of a user is introduced to resist the collusion attack of users. Subsequently, an anonymous key generation mechanism is equipped to prevent multiple AAs from building a full profile using the user's GID. (3) To ensure the correctness and integrity of EHRs, users can conduct double verification based on the verification tag and convergent key. Finally, the efficiency analysis and experiments show that the scheme meets the security requirements of key management and privacy preservation in cloud and is proven secure and efficient in practice under the decisional bilinear Diffie-Hellman (DBDH) assumption.

With the support of cloud computing, the scalability of the information-centric Internet of Things (IC-IoT) can be increased greatly. In the cloud-supported IC-IoT, how to make search more secure and efficient is a key issue to deal with. Searchable encryption is a potential technique to solve the above problem. In the cloud-supported IC-IoT, sensitive data are provided by multiple data users and can be accessed by specific data users. Then, the expression ability, search authorization, dynamic updates, and search efficiency become major concerns of search in the cloud-supported IC-IoT. In this paper, SASDQ, a semantic and authorized search scheme with dynamic updates and quick response, is proposed. In SASDQ, by constructing users’ personal dictionaries, dynamic updates and efficient index generation are achieved, and by leveraging ciphertext policy attribute-based encryption and conceptual graphs, semantic search with file-level fine-grained access authorization can be realized. In addition, quick search response is implemented by the preprocessing of time-consuming authorization tasks. The analysis shows that the proposed scheme can meet the data security requirement of the cloud-supported IC-IoT, and it also has less cost compared with other popular models.

Recommending items and friends are equally important in social recommendation applications. Current works usually consider them as two independent tasks and address them separately. This paper presents a new method to collaboratively perform the two tasks, in which three types of sparse data, ratings, trust, and distrust, are complemented with each other by a coupled low-rank approximation model. The data sparsity problems, which greatly decline prediction quality in respective tasks, are well mitigated in this way. The recovery conditions regarding the low-rank approximation are provided. Empirical studies on real data also support our theoretical models and analysis. To the best of our knowledge, this is the first attempt to raise and address such a problem.