Generating Questions Based on Semi-Automated and End-to-End Neural Network

With the emergence of large-scale knowledge base, how to use triple information to generate natural questions is a key technology in question answering systems. The traditional way of generating questions require a lot of manual intervention and produce lots of noise. To solve these problems, we propose a joint model based on semi-automated model and End-to-End neural network to automatically generate questions. The semi-automated model can generate question templates and real questions combining the knowledge base and center graph. The End-to-End neural network directly sends the knowledge base and real questions to BiLSTM network. Meanwhile, the attention mechanism is utilized in the decoding layer, which makes the triples and generated questions more relevant. Finally, the experimental results on SimpleQuestions demonstrate the effectiveness of the proposed approach.     

Relationship Between Shipping Accessibility and Maritime Transport Demand: the Case of Mainland China

Networks and Spatial Economics - The relationship between shipping accessibility and maritime transport demand is studied based on the relationship between production and consumption and stochastic...     

Long-term sustained-released in situ gels of a water-insoluble drug amphotericin B for mycotic arthritis intra-articular administration: preparation,in vitro and in vivo evaluation

Amphotericin B (AMB) was often used in intra-articular injection administration for fungal arthritis, because it could often bring a satisfactory therapeutic efficacy and a minimum systemic toxic side effect. However, because of the multiple operations and the frequent injections, the compliance of the patients was bad. Therefore, to develop a long-term sustained-released preparation of AMB for mycotic arthritis intra-articular administration is of great significance. The purpose of present study was to develop a long-term sustained-released in situ gel of a water-insoluble drug AMB for mycotic arthritis intra-articular administration. Based on the evaluations of the in vitro properties of the formulations, the formulation containing 10% (w/w) ethanol, 15% (w/w) PG, 0.75% (w/w) HA, 5% (w/w) purified soybean oil, 0.03% (w/w) α-tocopherol, 15% (w/w) water and 55% (w/w) glyceryl monooleate was selected as a suitable intra-articular injectable in situ gel drug delivery system for water-insoluble drug AMB. Furthermore, the results of the in vivo study on rabbits showed that the selected formulation was a safe and effective long-term sustained-released intra-articular injectable AMB preparation. Therefore, the presented in situ AMB gel could reduce the frequency of the administration in the AMB treatment of fungal arthritis, and then would get a good patient compliance.     

Nanomaterials for radiotherapeutics-based multimodal synergistic cancer therapy

Over the past decade, numerous studies have attempted to enhance the effectiveness of radiotherapy (external beam radiotherapy and internal radioisotope therapy) for cancer treatment. However, the low radiation absorption coefficient and radiation resistance of tumors remain major critical challenges for radiotherapy in the clinic. With the development of nanomedicine, nanomaterials in combination with radiotherapy offer the possibility to improve the efficiency of radiotherapy in tumors. Nanomaterials act not only as radiosensitizers to enhance radiation energy, but also as nanocarriers to deliver therapeutic units in combating radiation resistance. In this review, we discuss opportunities for a synergistic cancer therapy by combining radiotherapy based on nanomaterials designed for chemotherapy, photodynamic therapy, photothermal therapy, gas therapy, genetic therapy, and immunotherapy. We highlight how nanomaterials can be utilized to amplify antitumor radiation responses and describe cooperative enhancement interactions among these synergistic therapies. Moreover, the potential challenges and future prospects of radio-based nanomedicine to maximize their synergistic efficiency for cancer treatment are identified.

     

Ultrafast Electron–Phonon Coupling at Metal-Dielectric Interface

Energy transfer from photo-excited electrons in a metal thin film to the dielectric substrate is important for understanding the ultrafast heat transfer process across the two materials. Substantial research has been conducted to investigate heat transfer in a metal-dielectric structure. In this work, a two-temperature model in metal was used to analyze the interface electron and dielectric substrate coupling. An improved temperature and wavelength-dependent Drude–Lorentz model was implemented to interpret the signals obtained in optical measurements. Ultrafast pump-and-probe measurements on Au-Si samples were carried out, where the probe photon energy was chosen to be close to the interband transition threshold of gold to minimize the influence of non-equilibrium electrons on the optical response and maximize the thermal modulation to the optical reflectance. Electron-substrate interface thermal conductance at different pump laser fluences was obtained, and was found to increase with the interface temperature.     

Threshold privacy-preserving cloud auditing with multiple uploaders

International Journal of Information Security - Data integrity is a critical security issue in cloud storage. The data integrity checking schemes by a third-party auditor (TPA) have attracted a lot...     

Super-Resolution Reconstruction of Images Based on Microarray Camera

In the field of images and imaging, super-resolution (SR) reconstruction of images is a technique that converts one or more low-resolution (LR) images into a highresolution (HR) image. The classical two types of SR methods are mainly based on applying a single image or multiple images captured by a single camera. Microarray camera has the characteristics of small size, multi views, and the possibility of applying to portable devices. It has become a research hotspot in image processing. In this paper, we propose a SR reconstruction of images based on a microarray camera for sharpening and registration processing of array images. The array images are interpolated to obtain a HR image initially followed by a convolution neural network (CNN) procedure for enhancement. The convolution layers of our convolution neural network are 3×3 or 1×1 layers, of which the 1×1 layers are used to improve the network performance particularly. A bottleneck structure is applied to reduce the parameter numbers of the nonlinear mapping and to improve the nonlinear capability of the whole network. Finally, we use a 3×3 deconvolution layer to significantly reduce the number of parameters compared to the deconvolution layer of FSRCNN-s. The experiments show that the proposed method can not only ameliorate effectively the texture quality of the target image based on the array images information, but also further enhance the quality of the initial high resolution image by the improved CNN.