DEM: Deep Entity Matching Across Heterogeneous Information Networks

Heterogeneous information networks, which consist of multi-typed vertices representing objects and multi-typed edges representing relations between objects, are ubiquitous in the real world. In this paper, we study the problem of entity matching for heterogeneous information networks based on distributed network embedding and multi-layer perceptron with a highway network, and we propose a new method named DEM short for Deep Entity Matching. In contrast to the traditional entity matching methods, DEM utilizes the multi-layer perceptron with a highway network to explore the hidden relations to improve the performance of matching. Importantly, we incorporate DEM with the network embedding methodology, enabling highly efficient computing in a vectorized manner. DEM’s generic modeling of both the network structure and the entity attributes enables it to model various heterogeneous information networks flexibly. To illustrate its functionality, we apply the DEM algorithm to two real-world entity matching applications: user linkage under the social network analysis scenario that predicts the same or matched users in different social platforms and record linkage that predicts the same or matched records in different citation networks. Extensive experiments on real-world datasets demonstrate DEM’s effectiveness and rationality.

     

The Effect of Antibacterial Particle Incorporation on the Mechanical Properties,Biodegradability, and Biocompatibility of PLA and PHBV Composites

The composites based on polylactide (PLA) and poly (3‐hydroxybutyrate‐co‐3‐hydroxyvalerate) (PHBV) with the addition of antibacterial particles: silver (Ag) and copper oxide (CuO) are characterized. Basic mechanical properties and biodegradation processes, as well as biocompatibility of materials with human cells are determined. The addition of Ag or CuO to the polymers do not significantly affect their mechanical properties, flammability, or biodegradation rate. However, several differences between the base materials are observed. PLA‐based composites have higher tensile and impact strength values, while PHBV‐based ones have a higher modulus of elasticity, as well as better mechanical properties at elevated temperatures. Concerning biocompatibility, each of the tested materials support the growth of fibroblasts over time, although large differences are observed in the initial cell attachment. The analysis of hydrolytic degradation effects on the structure of materials shows that PHBV degrades much faster than PLA. The results of this study confirm the good potential of the investigated biodegradable polymer composites with antibacterial particles for future biomedical applications.     

基于深度残差网络的化工过程故障诊断

本工作提出了一种基于深度残差网络(DRN)的化工过程故障诊断方法，可从大量化工过程运行数据中自动提取故障特征。模型采用快捷连接缓解传统深度神经网络训练困难的问题，且使用批归一化(BN)方法，可有效缓解梯度消失/爆炸的问题。以田纳西?伊斯曼(TE)过程为实验对象对所提方法进行诊断性能评价实验，并与以往的基于传统深度学习模型的TE过程故障诊断方法进行对比，进一步探究了模型层数、BN技术和残差结构对故障诊断率的影响，最后，通过t分布随机邻域嵌入(t-SNE)方法对网络部分层的输出进行可视化。结果表明，模型对21种工况取得了94%的平均故障诊断率和0.30%的平均假阳率，表现出更加卓越的诊断性能。输出层的二维散点图显示了清晰的聚类，表明所提出的DRN模型能够对故障进行准确诊断。     

Detailed Insight into the Interaction of Bicyclic Somatostatin Analogue with Cu(II) Ions

Somatostatin analogues are useful pharmaceuticals in peptide receptor radionuclide therapy. In previous studies, we analyzed a new bicyclic somatostatin analogue (BCS) in connection with Cu(II) ions. Two characteristic sites were present in the peptide chain: the receptor- and the metal-binding site. We have already shown that this ligand can form very stable imidazole complexes with the metal ion. In this work, our aim was to characterize the intramolecular interaction that occurs in the peptide molecule. Therefore, we analyzed the coordination abilities of two cyclic ligands, i.e., P1 only with the metal binding site and P2 with both sites, but without the disulfide bond. Furthermore, we used magnetic circular dichroism (MCD) spectroscopy to better understand the coordination process. We applied this method to analyze spectra of P1, P2, and BCS, which we have described previously. Additionally, we analyzed the MCD spectra of P3 ligand, which has only the receptor binding site in its structure. We have unequivocally shown that the presence of the Phe-Trp-Lys-Thr motif and the disulfide bond significantly increases the metal binding efficiency.     

The Protective Effects of the Autophagic and Lysosomal Machinery in Vascular and Valvular Calcification: A Systematic Review

Background: Autophagy is a highly conserved catabolic homeostatic process, crucial for cell survival. It has been shown that autophagy can modulate different cardiovascular pathologies, including vascular calcification (VCN). Objective: To assess how modulation of autophagy, either through induction or inhibition, affects vascular and valvular calcification and to determine the therapeutic applicability of inducing autophagy. Data sources: A systematic review of English language articles using MEDLINE/PubMed, Web of Science (WoS) and the Cochrane library. The search terms included autophagy, autolysosome, mitophagy, endoplasmic reticulum (ER)-phagy, lysosomal, calcification and calcinosis. Study characteristics: Thirty-seven articles were selected based on pre-defined eligibility criteria. Thirty-three studies (89%) studied vascular smooth muscle cell (VSMC) calcification of which 27 (82%) studies investigated autophagy and six (18%) studies lysosomal function in VCN. Four studies (11%) studied aortic valve calcification (AVCN). Thirty-four studies were published in the time period 2015–2020 (92%). Conclusion: There is compelling evidence that both autophagy and lysosomal function are critical regulators of VCN, which opens new perspectives for treatment strategies. However, there are still challenges to overcome, such as the development of more selective pharmacological agents and standardization of methods to measure autophagic flux.     

Alternatives for improving energy security in Cape Verde

The lack of consensus on the meaning of energy security and how it should be measured hinders the evaluation of alternatives for improving the Cape Verdean energy sector. The objective of the paper is to develop an energy security index adjusted to the country's context. The methods used are the multi-criteria analysis and the Delphi survey. Different alternatives to improve energy security are assessed: electrically interconnect Cape Verde's islands, increase the share of renewable energy, and a combination of the previous alternatives. The results of our analysis indicate that increasing renewable energy is the optimal alternative for Cape Verde.     

分散式风电多点接入协调优化控制策略

针对分散式风电机组并网导致的配电网网损增加、电压稳定性降低等问题,文章提出一种分散式风电并网多点协调双层控制策略。首先,分析了分散式风电多点接入特性,研究分散式风电多点接入后的网损、电压分布和功率因数的关系,并提出最小网损和电压偏差为目标的双层优化控制策略。最后,基于IEEE-33节点进行仿真计算,结果表明,所提多点协调控制策略能够有效降低配电网网损和提升母线电压水平,增加系统的稳定性。与超前功率因数相比,滞后的功率因数能够增大风能消纳能力,提高供电可靠性,减少电压偏移水平。     

基于Adam-DBN网络的行星齿轮箱故障诊断方法研究

随着现代机械装备的复杂化,传统的故障诊断方法难以满足表征设备间的复杂映射关系,且在如今大数据背景下面临着维数灾难的问题。文章结合深度置信网络以及梯度优化算法提出了一种基于梯度优化深度置信网络(Adam-DBN),通过数据验证选取最优梯度优化算法对深度置信网络的的梯度算法进行调优。搭建模拟实际工况的行星齿轮箱实验台,通过实验台采集所得数据构成数据集对方法进行验证。实验表明文中提出的方法能够有效提高DBN网络的收敛速度与训练精度,同时具有较高的故障识别准确率。