Deep Multilabel Multilingual Document Learning for Cross-Lingual Document Retrieval

Cross-lingual document retrieval, which aims to take a query in one language to retrieve relevant documents in another, has attracted strong research interest in the last decades. Most studies on this task start with cross-lingual comparisons at the word level and then represent documents via word embeddings, which leads to insufficient structure information. In this work, the cross-lingual comparison at the document level is achieved through the cross-lingual semantic space. Our method, MDL (deep multilabel multilingual document learning), leverages a six-layer fully connected network to project cross-lingual documents into a shared semantic space. The semantic distances can be calculated when the cross-lingual documents are transformed into embeddings in semantic space. The supervision signals are automatically extracted from the data and then used to construct the semantic space via a linear classifier. The ambiguity of manual labels could be avoided and the multilabel supervision signals can be acquired instead of a single label. The representation of the semantic space is enriched by multilabel supervision signals, which improves the discriminative ability of the embeddings. The MDL is easy to extend to other fields since it does not depend on specific data. Furthermore, MDL is more efficient than the models training all languages jointly, since each language is trained individually. Experiments on Wikipedia data showed that the proposed method outperforms the state-of-the-art cross-lingual document retrieval methods.     

聚合物气体辅助注射成型制品的形态结构

作为一种由常规注射成型发展起来的聚合物加工技术,气体辅助注射成型具有节约原材料、缩短成型周期以及提高制品性能等优点,已得到广泛的应用.由于气辅成型过程是一个在刚、柔双重约束界面条件下进行的多相复杂体系的多次流动过程,因而其形态结构的形成、发展和演化要远比常规注塑成型复杂.然而在气辅成型的形态结构方面,国内外的研究一直以来开展得较少.近年来,作者在聚合物及其共混物、复合材料气辅成型制品的形态结构方面已开展了广泛的研究工作,本文对这些工作和一些重要结果作了总结,并简要分析了成型过程中剪切场对形态演化的影响,最后对该研究方向的发展趋势作了展望.     

Tailoring Pore Structure and Morphologies in Covalent Organic Frameworks for Xe/Kr Capture and Separation

As a rising star among porous solid materials, covalent organic frameworks(COFs) with excellent properties including but not limit to facilely controllable structure, high porosity, and multi-chemical functionality represent significant potential for efficient ¹²⁷Xe/⁸⁵Kr capture and separation. In this study, through tuning the length of the organic ligands, two-dimensional(2D) COF mate-rials with identical connection group but different pore properties, denoted as ATFG-COF and TpPa-COF with AA-stacking model and TpBD-COF with AB-stacking model were synthesized and tested for Kr and Xe adsorption for the first time. Adsorption measurements indicate that the narrower pore apertures and higher porosity are conducive for COF materials to capture Xe and Kr. Furthermore, the Henry's constant, isosteric heat of adsorption(Q_st), and ideal adsorbed solution theory(IAST) selectivity of ATFG-COF, the pore size of which is closest to the kinetic diameter of the Xe atom(0.41 nm) among 2D COF materials, were carried out based on the single component sorption isotherms. The results illustrate that the high isosteric heat values of Xe/Kr adsorption on ATFG-COF are 25 and 16 kJ/mol at room temperature, respectively. Henry's law predicts that the selectivity factor of Xe to Kr is 6.07, consistent with the adsorption selectivity(ca. 6) calculated based on the IAST.     

Genetically engineered materials: Proteins and beyond

Information-rich molecules provide opportunities for evolution. Genetically engineered materials are superior in that their properties are coded within genetic sequences and could be fine-tuned. In this review, we elaborate the concept of genetically engineered materials(GEMs) using examples ranging from engineered protein materials to engineered living materials. Proteinbased materials are the materials of choice by nature. Recent progress in protein engineering has led to opportunities to tune...     

嵌段共聚物自组装的研究进展

嵌段共聚物自组装在光学、电子、信息、化学及生物领域有着广泛的应用前景.本文从实验观测、理论研究和计算机模拟三个方面概述了嵌段共聚物自组装领域的研究进展.在实验观测方面,着重介绍了嵌段共聚物在体相及膜中的自组装及外场调控作用方面的研究进展;理论方面则分别介绍了强分相理论、弱分相理论、自洽场理论、动态密度泛函方法和元胞动力学等在嵌段共聚物自组装领域的应用;计算机模拟方面就 Monte Carlo 模拟、耗散粒子动力学等方法在该领域的应用作了详细的阐述.     

An Improvement to Conformer-Based Model for High-Accuracy Speech Feature Extraction and Learning

Owing to the loss of effective information and incomplete feature extraction caused by the convolution and pooling operations in a convolution subsampling network, the accuracy and speed of current speech processing architectures based on the conformer model are influenced because the shallow features of speech signals are not completely extracted. To solve these problems, in this study, we researched a method that used a capsule network to improve the accuracy of feature extraction in a conformer-based model, and then, we proposed a new end-to-end model architecture for speech recognition. First, to improve the accuracy of speech feature extraction, a capsule network with a dynamic routing mechanism was introduced into the conformer model; thus, the structural information in speech was preserved, and it was input to the conformer blocks via sequestered vectors; the learning ability of the conformed-based model was significantly enhanced using dynamic weight updating. Second, a residual network was added to the capsule blocks, thus, the mapping ability of our model was improved and the training difficulty was reduced. Furthermore, the bi-transformer model was adopted in the decoding network to promote the consistency of the hypotheses in different directions through bidirectional modeling. Finally, the effectiveness and robustness of the proposed model were verified against different types of recognition models by performing multiple sets of experiments. The experimental results demonstrated that our speech recognition model achieved a lower word error rate without a language model because of the higher accuracy of speech feature extraction and learning using our model architecture with a capsule network. Furthermore, our model architecture benefited from the advantage of the capsule network and the conformer encoder, and also has potential for other speech-related applications.     

Cloning,Expression, Purification,and Characterization of β-Galactosidase from Bifidobacterium longum and Bifidobacterium pseudocatenulatum

Expression and purification of β-galactosidases derived from Bifidobacterium provide a new resource for efficient lactose hydrolysis and lactose intolerance alleviation. Here, we cloned and expressed two β-galactosidases derived from Bifidobacterium. The optimal pH for BLGLB1 was 5.5, and the optimal temperature was 45 °C, at which the enzyme activity of BLGLB1 was higher than that of commercial enzyme E (300 ± 3.6 U/mg) under its optimal conditions, reaching 2200 ± 15 U/mg. The optimal pH and temperature for BPGLB1 were 6.0 and 45 °C, respectively, and the enzyme activity (0.58 ± 0.03 U/mg) under optimum conditions was significantly lower than that of BLGLB1. The structures of the two β-galactosidase were similar, with all known key sites conserved. When o-nitrophenyl-β-D-galactoside (oNPG) was used as an enzyme reaction substrate, the maximum reaction velocity (V_max) for BLGLB1 and BPGLB1 was 3700 ± 100 U/mg and 1.1 ± 0.1 U/mg, respectively. The kinetic constant (K_m) of BLGLB1 and BPGLB1 was 1.9 ± 0.1 and 1.3 ± 0.3 mmol/L, respectively. The respective catalytic constant (k_cat) of BLGLB1 and BPGLB1 was 1700 ± 40 s⁻¹ and 0.5 ± 0.02 s⁻¹, respectively; the respective k_cat/K_m value of BLGLB1 and BPGLB1 was 870 L/(mmol∙s) and 0.36 L/(mmol∙s), respectively. The K_m, k_cat and V_max values of BLGLB1 were superior to those of earlier reported β-galactosidase derived from Bifidobacterium. Overall, BLGLB1 has potential application in the food industry.     

林杏琴会堂的音质分析及改进设想

采用精密声级计和双通道声学分析仪，对林杏琴会堂的声场分布、噪声本底和混响时间作了分区测试，应用建筑声学理论对会堂的音质状况进行了分析，并提出改善会堂音质的若干建议．     

呋喃唑酮在凡纳滨对虾组织中代谢动力学研究

运用液相色谱-串联质谱法研究了呋喃唑酮药饵多次给药后在凡纳滨对虾体内的药物代谢动力学.研究结果表明：呋喃唑酮代谢物在血淋巴、肝胰脏、肌肉组织中的代谢过程均符合二室模型,其动力学方程分别为C血液=414.107×e^-0.092t＋124.451×e^-0.005t,C肝胰腺=1625.563×e^-0.019t＋125.700,C肌肉=120.434×e^-0.019t＋71.579×e^-0.001t.凡纳滨对虾血淋巴和肝胰腺中呋喃唑酮代谢物浓度在4h达到高峰,肌肉中药物浓度在2h达到最大值;最高峰时血淋巴、肝胰腺和肌肉中药物浓度平均为570.6μg·L^-1、1948.6μg·kg-1和210.0μg·kg-1.在血淋巴、肝胰腺和肌肉组织中呋喃唑酮代谢物浓度的吸收半衰期（t1/2α）分别为7.497h、36.206h和36.484h,消除半衰期（t1/2β）分别为132.525 h、1 000 479.217 h和505.637 h,总表观分布容积（V1/F）分别为55.775 L·kg^-1、17.16 L·kg^-1和161.574L·kg^-1.说明给药后呋喃唑酮代谢物在凡纳滨对虾体内消除缓慢,残留严重,尤其以肝胰脏残留最为明显.     

关联规则增量式更新的改进算法

关联规则是数据挖掘中的重要研究内容之一，对在事务数据库不变，只对最小支持度和最小可信度进行改变的情况下，针对关联规则的维护问题，设计了一个增量式更新的改进算法AIUA。