Enhanced Extraction Efficiency of Flavonoids from Pyrus ussuriensis Leaves with Deep Eutectic Solvents

In this study, deep eutectic solvents (DESs) were synthesized using different ratios of choline chloride (CC) and dicarboxylic acids, and their eutectic temperatures were determined. The DES synthesized using CC and glutaric acid (GA), which showed a higher extraction efficiency than conventional solvents, was used for the extraction of flavonoid components from Pyrus ussuriensis leaves (PUL), and the extraction efficiency was evaluated using the response surface methodology. The flavonoid components rutin, hyperoside, and isoquercitrin were identified through high-performance liquid chromatography (HPLC), equipped with a Waters 2996 PDA detector, and HPLC mass spectrometry (LC-MS/MS) analyses. The optimum extraction was achieved at a temperature of 30 °C using DES in a concentration of 30.85 wt.% at a stirring speed of 1113 rpm and an extraction time of 1 h. The corresponding flavonoid content was 217.56 μg/mL. The results were verified by performing three reproducibility experiments, and a high significance, with a confidence range of 95%, was achieved. In addition, the PUL extracts exhibited appreciable antioxidant activity. The results showed that the extraction process using the DES based on CC and GA in a 1:1 molar ratio could effectively improve the yield of flavonoids from PUL.     

Deep Learning-Based Method for Compound Identification in NMR Spectra of Mixtures

Nuclear magnetic resonance (NMR) spectroscopy is highly unbiased and reproducible, which provides us a powerful tool to analyze mixtures consisting of small molecules. However, the compound identification in NMR spectra of mixtures is highly challenging because of chemical shift variations of the same compound in different mixtures and peak overlapping among molecules. Here, we present a pseudo-Siamese convolutional neural network method (pSCNN) to identify compounds in mixtures for NMR spectroscopy. A data augmentation method was implemented for the superposition of several NMR spectra sampled from a spectral database with random noises. The augmented dataset was split and used to train, validate and test the pSCNN model. Two experimental NMR datasets (flavor mixtures and additional flavor mixture) were acquired to benchmark its performance in real applications. The results show that the proposed method can achieve good performances in the augmented test set (ACC = 99.80%, TPR = 99.70% and FPR = 0.10%), the flavor mixtures dataset (ACC = 97.62%, TPR = 96.44% and FPR = 2.29%) and the additional flavor mixture dataset (ACC = 91.67%, TPR = 100.00% and FPR = 10.53%). We have demonstrated that the translational invariance of convolutional neural networks can solve the chemical shift variation problem in NMR spectra. In summary, pSCNN is an off-the-shelf method to identify compounds in mixtures for NMR spectroscopy because of its accuracy in compound identification and robustness to chemical shift variation.     

青藏高原对北大西洋深水形成影响机制的季节差异

利用耦合地球系统模式(CESM1.0),通过对比有青藏高原的控制试验和无青藏高原的敏感性试验,定性地分析青藏高原对不同季节北大西洋深水(NADW)形成的影响机制.研究结果表明,青藏高原对NADW形成的影响机制因季节而异.移除青藏高原后,NADW形成将会减弱,冷季(北半球10月—次年3月)NADW形成的显著减弱是由海洋表...     

基于BP神经网络和Logistic模型的中国人均粮食产量预测研究

文章提出了一种基于长短期记忆网络(Long-short Term Memory network, LSTM)的两阶段均值-CVaR投资组合模型(LSTM+CVaR)。该模型在第一阶段采用LSTM预测股票收益并对股票进行选择;在第二阶段运用均值-CVaR模型来确定所选股票的投资比例。最后,以沪深300指数股为样本数据,在考虑交易成本和上界约束的情况下,比较LSTM+CVaR模型、LSTM预测选股的等比例模型、随机选股的CVaR模型、随机选股的等比例模型和沪深300指数的风险收益特征、累计收益率和夏普比率。实证结果表明：LSTM+CVaR模型能够实现比传统的投资组合模型更高的平均收益率、收益风险比、累计收益率和夏普比率;减少交易成本和放宽上界约束能提升投资组合模型的表现。

     

一种基于自注意力机制的文本图像生成对抗网络

图像自动生成一直以来都是计算机视觉领域的一项重要挑战,其中的文本到图像的生成更是图像生成领域的重要分支。随着深度学习技术迅猛发展,生成对抗网络的出现使得图像生成领域焕发生机,借助生成对抗网络能够生成较为生动且多样的图像。本文将自注意力机制引入生成对抗网络,提出GAN-SelfAtt以提升生成图像的质量。同时,使用WGAN、WGAN-GP 2种生成对抗网络框架对GAN-SelfAtt进行实现。实验结果表明,自注意力机制的引入能够提高生成图像的清晰度,这归功于自注意力机制弥补了卷积运算中只能计算局部像素区域内的相关性的缺陷。除此之外,GAN-SelfAtt在训练时有着更好的稳定性,避免了原始生成对抗网络中的模式坍塌问题。     

Highly Efficient Low Melting Mixture Catalyzed Synthesis of 1,8-Dioxo-dodecahydroxanthene Derivatives

A low melting citric acid‐urea mixture was found to be a new and effective promoting medium for the environmentally friendly synthesis of 1,8‐dioxo‐dodecahydroxanthene derivatives by condensation of aromatic aldehydes and cyclic 1,3‐dicarbobonyl compounds. This deep eutectic solvent acts as both the reaction medium and catalyst, furnishing xanthenediones in high to excellent yields.     

Hydrated‐Metal‐Halide‐Based Deep‐Eutectic‐Solvent‐Mediated NiFe Layered Double Hydroxide: An Excellent Electrocatalyst for Urea Electrolysis and Water Splitting

The liquid structures of deep eutectic solvents (DESs) based on hydrated metal halides and their application as electrolytes have been widely studied. However, little attention has been paid to the direct use of this type of DES in the preparation of micro‐/nanomaterials. Herein, an FeCl₃ ? 6 H₂O/urea DES was used in the one‐step synthesis of NiFe‐LDH_D with a nanoflower morphology. In alkaline media, this catalyst promoted excellent electrocatalytic activity for the oxidation of urea at potential of 1.32 V (vs. RHE) and for the oxygen‐evolution reaction at a potential of 1.39 V to achieve a current density of 10 mA cm^?2. These results were superior to the results with NiFe‐LDH/NF that was obtained from an aqueous solution of FeCl₃, as well as most of the previously reported transition‐metal catalysts. Furthermore, NiFe‐LDH_D/NF could be readily implemented as both a cathode and an anode for the electrolysis of urea and water splitting. The use of hydrated‐metal‐halide‐based DESs for the preparation of LDH catalysts through a dipping‐redox strategy should both enrich the research of DESs and offer guidance for the rational surface engineering of catalysts for the electrolysis of urea and overall water splitting with high performance.     

Deep Oxidation of Methane Over Cu and Ag Modified Manganese Oxide Catalysts

Cu-Mn and Ag-Mn composite oxide catalysts were prepared by the modification of high specific surface area MnO₂ precursor with Cu or Ag, and used for CH₄ deep oxidation. The results were compared with that of unmodified MnOx. It is found that both Cu and Ag additives increase the oxidation activity of manganese oxide below 420°C; however, only Ag lowers the CH₄ complete oxidation temperature (T₉₈), whereas Cu increases the T₉₈. The catalysts were characterized by means of BET, XRD and H₂-TPR. Due to the synergetic effects of Cu and Ag ions with manganese oxide, the activity of the oxygen species is improved, leading to the increase of the oxidation activities of the modified catalysts, especially below 420°C.     

基于排名学习和多源信息的地图匹配方法

融合多源信息能有效提高地图匹配的准确率。已有的地图匹配方法依赖于数学模型,当引入新类型的数据时, 需要重新设计数学模型或调整模型参数。为解决该问题,提出了一种端到端的数据驱动地图匹配方法。该方法不需要建立具体的数学模型,只需从匹配结果已知的数据中学习候选道路的评分函数：选出某GPS点的候选道路,利用评分函数对所有候选道路进行打分,选择分数最高的道路作为地图匹配结果。实验结果表明,该方法能直接利用新类型的数据提高地图匹配的准确率,能在数据缺失时避免准确率急剧降低。此外,具有与基于HMM方法相近的准确率和与基于夹角特征和距离特征方法相当的速度。     

青藏高原日土地区构造单元划分,特征及演化

西藏高原的构造单元及深断裂基本上汇聚在日土地区,是揭开青藏高原成因奥秘关键之地。笔者将该地区划分为6个亚一级和若干次级构造单元:昆仑海西、冈底斯早燕山两褶皱系和巴颜喀拉印支褶皱带具典型地槽性质,喀喇昆仑晚海西、喜马拉雅晚燕山两褶皱系和羌塘地块趋于准地台性质。各单元由活动转稳定过程、深断裂形成与封闭、岩浆活动等均具北老南新逐渐阶段性迁移特征;岩石圈固结有自北而南逐渐推移扩大趋势。巨大蛇绿岩带显示新生地槽的发展演化。