Att-U-Net：融合注意力机制的U-Net骨导语声增强*

近年来大量全卷积网络、U-Net等编解码网络结构应用于语音增强,它们具有计算复杂度低、模型参数少等优势。然而,与长短时记忆模型等方法相比,这些编解码结构仍存在不能充分利用先后时间之间和高低频率之间的关联信息等缺点,尤其对于长序列数据的输入,编解码结构存在信息丢失的问题。为保持计算效率的同时考虑更充分的时频关联信息建模,本文提出一种融合注意力机制的U-Net网络的骨导语音增强方法（Att-U-Net）,通过在跳跃连接中引入注意力机制,生成一个权重矩阵,将编码层中的全局信息根据权重融入对应的解码层中,使网络在编解码过程中能够关注输入数据中与增强目标相关程度高的重要信息,同时抑制不相关的信息。在骨导语音数据集上的实验表明,融合注意力机制的U-Net网络能在保持模型轻量化的同时有效提升骨导语音的增强效果,增强后的语音在各项客观评价指标上均优于基线模型。通过对编解码网络中间层的可视化分析发现,在解码过程中注意力机制有效地保留了有声段的信息,滤除了骨导语音由于骨导传声特性带来的中频共振,从而使得增强后的骨导语音具有较好的听觉效果。     

Novel superabsorbent polymer composites based on α-cellulose and modified zeolite: synthesis,characterization, water absorbency and water retention capacity

Cellulose - Most superabsorbent polymers (SAPs) are prepared based on synthetic polymers (from petroleum resources), making them costly, nondegradable, and not ecofriendly. To overcome these...     

近红外光谱小波分析在土壤参数预测中的应用

从田间采集了150个田间土壤样本,在分析了所有样本的土壤参数统计特征之后,对原始近红外光谱数据进行了聚类分析,分别得到了50个土壤全氮和50个土壤有机质的等价样本及其对应光谱。对样本光谱曲线进行8层Biorthogonal小波包分解,分解得到的最低低频[80]结点对应着土壤水分以及土壤质地的光谱变化趋势,最高高频[8 255]结点对应着土壤粒度、光谱仪精度等引起的高频震荡。对以上两个结点进行重构并从样本光谱曲线中剔除以上影响成分,得到了对应的土壤参数特征光谱。基于特征光谱建立了土壤参数偏最小二乘回归模型：全氮偏最小二乘预测模型的预测系数r_c达到了0.960,验证系数r_v达到了0.920;有机质偏最小二乘预测模型的预测系数r_c达到0.922,验证系数r_v达到0.883。模型精度明显提高,满足实际生产的需要。     

Prediction of quantitative calibration factors of some organic compounds in gas chromatography

A quantitative structure-property relationship (QSPR) methodology that involves multilinear (Hansch-type) and nonlinear (radial basis function neural network (RBFNN)) approaches was performed to correlate the quantitative molar calibration factors (f(M)) of 140 organic compounds against structural factors. The statistical characteristics provided by the multiple linear model (R(2) = 0.963; RMS = 0.089; AARD = 3.86% for test set) indicated satisfactory stability and predictive ability, while the predictive ability of the RBFNN model is somewhat superior (R(2) = 0.983; RMS = 0.075; AARD = 3.19% for test set). The multilinear model provided some insight into the main structure factors that modulate the quantitative calibration factor of the investigated compounds.     

An Unlocked Two-Dimensional Conductive Zn-MOF on Polymeric Carbon Nitride for Photocatalytic H2O2 Production

Developing highly efficient catalytic sites for O₂ reduction to H₂O₂, while ensuring the fast injection of energetic electrons into these sites, is crucial for artificial H₂O₂ photosynthesis but remains challenging. Herein, we report a strongly coupled hybrid photocatalyst comprising polymeric carbon nitride (CN) and a two-dimensional conductive Zn-containing metal–organic framework (Zn-MOF) (denoted as CN/Zn-MOF(lc)/400; lc, low crystallinity; 400, annealing temperature in °C), in which the catalytic capability of Zn-MOF(lc) for H₂O₂ production is unlocked by the annealing-induced effects. As revealed by experimental and theoretical calculation results, the Zn sites coordinated to four O (Zn-O₄) in Zn-MOF(lc) are thermally activated to a relatively electron-rich state due to the annealing-induced local structure shrinkage, which favors the formation of a key *OOH intermediate of 2e⁻ O₂ reduction on these sites. Moreover, the annealing treatment facilitates the photoelectron migration from the CN photocatalyst to the Zn-MOF(lc) catalytic unit. As a result, the optimized catalyst exhibits dramatically enhanced H₂O₂ production activity and excellent stability under visible light irradiation.     

Insight into 6π Electrocyclic Reactions of 1,8‐Dioxatetraene

A variety of benzofuranone‐based spiroisochromenes were originally designed and synthesized to gain insight into the oxa‐6π electrocyclic reaction of cis,cis‐1,8‐dioxatetraene for the first time. The stability of the 1,8‐dioxatetraene intermediate is governed by its steric congestion and can be fine‐tuned through modification of the backbone structure, leading to the reactivity differences in the 6π electrocyclic reaction and the emergence of photochromic properties.     

十二烷基甜菜碱/十二烷基硫酸钠复配体系的表面活性

研究了两性表面活性剂十二烷基甜菜碱（C12BE）和阴离子型表面活性剂十二烷基硫酸钠（SDS）复配体系的形成胶束能力,降低表面张力效率,降低表面张力能力三种增效作用.发现C12BE和SDS摩尔比为6 :4时增效作用最显著.并考察了盐、醇对复配体系表面活性的影响,结果表明,加盐能导致表面活性的提高,加醇的机理比较复杂,乙醇对复配体系表面活性影响不大,而正丁醇影响则比较显著.     

发酵法黄腐酸磺化产物减水性能的研究

以腐植酸（HA）类新技术产品－发酵法黄腐酸（BFA）的磺化（磺甲基化）产物为研究对象,就其物化特性和泡沫性能以及表面活性、水泥净浆流动度和减水率等一系列减水性能进行了较系统的考察和研究,并与常用减水剂木钙（木质磺酸钙）对比,提出发酵法黄腐酸应用于混凝土减水剂的可能性。文中还对BFA磺化（磺甲基化）产物的减水机理进行了探讨。     

巯基乙酸锑(Ⅲ)配合物合成与晶体结构

以三氧化二锑和巯基乙酸在水溶液中反应合成了配合物HSb(SCH2COO)2,并通过元素分析、红外光谱、X射线粉末衍射进行了表征,利用单晶X射线四圆衍射法测定了晶体结构,结果表明该配合物晶体属于单斜晶系,C2/c空间群,晶胞参数为:a=1.40005(8)nm,b=1.19121(8)nm,c=1.23588(8)nm,β=126.822(1)°,V=1.6499(2)nm^3,Dc=2.439g·cm^-3,Z=8,R1=0.0250。并对X射线粉末衍射数据进行了指标化,其结果与单晶数据吻合。     

Oriented Antibody Immobilization and Immunoassay Based on Boronic Acid-containing Polymer Brush

High sensitive immunoassay platforms have gained intense attention due to their vital roles in early-stage disease diagnosis and therapeutic information feedback.Although random covalent-binding of antibody has been widely adopted in immunoassays due to its simplicity and effectiveness,it readily loses its activity and fails to exhibit high antigen-binding capacity.In this work,copolymer of zwitterionic sulfobetaine methacrylate (SBMA) and glycidyl methacrylate (GMA) brushes were immobilized onto inert polypropylene (PP) via surface-initiated atom transfer radical polymerization (ATRP) based on biomimetic dopamine pretreatment.Subsequently,boronic acid (BA) groups were covalently bonded via active GMA units,followed by the introduction of oriented immobilization of antibody.Owing to the oriented immobilization of antibody facilitated by BA groups in polymer brush,the bioactivity of antibody is well preserved,which endows the surface with significantly enhanced antigen-binding capacity.Moreover,the existence of SBMA segments in polymer brushes renders the surface high resistance to nonspecific protein adsorption,significantly alleviating the signal interference of antigen recognition.This strategy could find potential applications in developing high sensitive immunoassay platforms based on the different substrates.