Neural Machine Translation by Fusing Key Information of Text

When the Transformer proposed by Google in 2017, it was first used for machine translation tasks and achieved the state of the art at that time. Although the current neural machine translation model can generate high quality translation results, there are still mistranslations and omissions in the translation of key information of long sentences. On the other hand, the most important part in traditional translation tasks is the translation of key information. In the translation results, as long as the key information is translated accurately and completely, even if other parts of the results are translated incorrect, the final translation results’ quality can still be guaranteed. In order to solve the problem of mistranslation and missed translation effectively, and improve the accuracy and completeness of long sentence translation in machine translation, this paper proposes a key information fused neural machine translation model based on Transformer. The model proposed in this paper extracts the keywords of the source language text separately as the input of the encoder. After the same encoding as the source language text, it is fused with the output of the source language text encoded by the encoder, then the key information is processed and input into the decoder. With incorporating keyword information from the source language sentence, the model’s performance in the task of translating long sentences is very reliable. In order to verify the effectiveness of the method of fusion of key information proposed in this paper, a series of experiments were carried out on the verification set. The experimental results show that the Bilingual Evaluation Understudy (BLEU) score of the model proposed in this paper on the Workshop on Machine Translation (WMT) 2017 test dataset is higher than the BLEU score of Transformer proposed by Google on the WMT2017 test dataset. The experimental results show the advantages of the model proposed in this paper.     

Regulation of Polymer Configurations Enables Green Solvent-Processed Large-Area Binary All-Polymer Solar Cells With Breakthrough Performance and High Efficiency Stretchability Factor

With the great potential of the all-polymer solar cells for large-area wearable devices, both large-area device efficiency and mechanical flexibility are very critical but attract limited attention. In this work, from the perspective of the polymer configurations, two types of terpolymer acceptors PYTX-A and PYTX-B (X = Cl or H) are developed. The configuration difference caused by the replacement of non-conjugated units results in distinct photovoltaic performance and mechanical flexibility. Benefiting from a good match between the intrinsically slow film-forming of the active materials and the technically slow film-forming of the blade-coating process, the toluene-processed large-area (1.21 cm²) binary device achieves a record efficiency of 14.70%. More importantly, a new parameter of efficiency stretchability factor (ESF) is proposed for the first time to comprehensively evaluate the overall device performance. PM6:PYTCl-A and PM6:PYTCl-B yield significantly higher ESF than PM6:PY-IT. Further blending with non-conjugated polymer donor PM6-A, the best ESF of 3.12% is achieved for PM6-A:PYTCl-A, which is among the highest comprehensive performances.     

Influence of the silica matrix on the formation of α-alumina in a mullite–alumina composite from a diphasic precursor

In mullite–alumina composite precursors, interaction between the silica matrix and the fine -alumina texture strongly influences the precursor phase transformation, the nucleation and the crystal geometry both of the mullite and of the -alumina. The mullite–alumina composite precursor calcined at 1000 °C has a layered structure probably derived from the layered texture of the -alumina. The phase transition of this layered texture is retarded by the presence of the silica matrix and a metastable mullite phase is formed before nucleation of -alumina. By leaching away the silica matrix, the remaining layered texture is readily transformed into very fine, thin -alumina platelets by calcination at 1000 °C. This seems to be one reason for the appearance of elongated mullite grains in a pure mullite matrix and the platelet shaped -alumina grains in the mullite–alumina composite prepared from diphasic precursors. © 1998 Chapman & Hall     

Effect of axial and radial velocities on solids mass flow rate measurement

In coal powder flow transportation and combustion, powder mass flow rate is a key parameter to be monitored and controlled. Electrostatics is one of the techniques used for such task with its non-intrusive, robust, and low cost natures. The passive electrostatic meters measure charge induced on the detecting electrodes. As it is known that the induced signal is not only dependent on the solids mass flow rate, but also affected by solids velocity. However, the velocity of particles usually referred to is the axial velocity. In reality, the solids velocity is a vector, its projections in both the radial and tangential directions also need to be investigated. This paper analyses the dynamic sensitivity of ring-shaped electrostatic sensors using the finite element method (FEM), and investigates the influence of the axial and radial velocities on the induced signal on the electrodes.     

Array SSL VPN如何保障番禺有线办公网的安全运行

企业办公自动化系统的应用十分广泛,随着互联网的高速发展,WEB应用成为主流,OA系统在互联网上的延伸成为大趋势,本文阐述在WEB应用中如何保障数据和系统安全的手段和方法。     

Stabilization of linear systems with input delay and saturation—A parametric Lyapunov equation approach

This paper studies the problem of stabilizing a linear system with delayed and saturating feedback. It is known that the eigenstructure assignment‐based low‐gain feedback law (globally) stabilizes a linear system in the presence of arbitrarily large delay in its input, and semi‐globally stabilizes it when the input is also subject to saturation, as long as all its open‐loop poles are located in the closed left‐half plane. Based on a recently developed parametric Lyapunov equation‐based low‐gain feedback design method, this paper presents alternative, but simpler and more elegant, feedback laws that solve these problems. The advantages of this new approach include its simplicity, the capability of giving explicit conditions to guarantee the stability of the closed‐loop system, and the ease in scheduling the low‐gain parameter on line to achieve global stabilization in the presence of actuator saturation. Copyright © 2009 John Wiley & Sons, Ltd.     

崔岫闻：与自己合谋一场革命

对于已经功成名就的艺术家而言,最难摆脱的可能就是在被认知后对于自我体系的不断重复。然而这也正是艺术的魅力所在,超越与重生,以崭新的姿态面对世界,这样的探索,本身就是一种涅磐。