基于GA_XGBoost模型的大学生科研能力预测问题研究

科学评价大学生科研创新能力对我国科研水平的提高具有重要意义.采用机器学习模型来预测大学生科研能力可以起到良好的效果,提出一种GAXGBoost模型来实现对大学生的科研能力预测.此模型是以Xgboost算法为基础,然后充分利用遗传算法的全局搜索能力自动搜索Xgboost最优超参数,避免了人为经验调参不准确的缺陷,最后采用精英选择策略以此确保每一轮都是最佳的进化结果.通过分析表明,所采用的GAXGBoost模型在大学生科研能力预测的结果中具有很高的精度,将此模型与Logistic Regression、Random Forest、SVM等模型进行对比,GAXGBoost模型的预测精度最高.     

Transfer learning in deep neural network based under-sampled MR image reconstruction

In Magnetic Resonance Imaging (MRI), the success of deep learning-based under-sampled MR image reconstruction depends on: (i) size of the training dataset, (ii) generalization capabilities of the trained neural network. Whenever there is a mismatch between the training and testing data, there is a need to retrain the neural network from scratch with thousands of MR images obtained using the same protocol. This may not be possible in MRI as it is costly and time consuming to acquire data. In this research, a transfer learning approach i.e. end-to-end fine tuning is proposed for U-Net to address the data scarcity and generalization problems of deep learning-based MR image reconstruction. First the generalization capabilities of a pre-trained U-Net (initially trained on the human brain images of 1.5 T scanner) are assessed for: (a) MR images acquired from MRI scanners of different magnetic field strengths, (b) MR images of different anatomies and (c) MR images under-sampled by different acceleration factors. Later, end-to-end fine tuning of the pre-trained U-Net is proposed for the reconstruction of the above-mentioned MR images (i.e. (a), (b) and (c)). The results show successful reconstructions obtained from the proposed method as reflected by the Structural SIMilarity index, Root Mean Square Error, Peak Signal-to-Noise Ratio and central line profile of the reconstructed images.     

Distribution dependent stochastic differential equations

Due to their intrinsic link with nonlinear Fokker-Planck equations and many other applications, distribution dependent stochastic differential equations (DDSDEs) have been intensively investigated. In this paper, we summarize some recent progresses in the study of DDSDEs, which include the correspondence of weak solutions and nonlinear Fokker-Planck equations, the well-posedness, regularity estimates, exponential ergodicity, long time large deviations, and comparison theorems.     

Embedded exponential-type low-regularity integrators for KdV equation under rough data

BIT Numerical Mathematics - In this paper, we introduce a novel class of embedded exponential-type low-regularity integrators (ELRIs) for solving the KdV equation and establish their optimal...     

[Nb@Ge13/14]3−: New Family Members of Ge-Based Intermetalloid Clusters

During the past two decades, single-atom-centered medium-sized germanium clusters [M@Ge_n] (M=transition metals, n>12) have been extensively explored, both from theoretical perspectives and experimental gas-phase syntheses. However, the actual structural arrangements of the Ge₁₃ and Ge₁₄ endohedral cages are still ambiguous and have long remained an unresolved problem for experimental implementation. In this work, we successfully synthesize 13-/14-vertex Ge clusters [Nb@Ge₁₃]³⁻ ( 1 ) and [Nb@Ge₁₄]³⁻ ( 2 ), which are structurally characterized and exhibit unprecedented topologies, neither classical deltahedra nor 3-connected polyhedral structures. Theoretical analysis indicates that the major stabilization of the Ge backbones arises due to the substantial interaction of Ge 4p-AOs with the endohedral Nb 4d-AOs through three/four-center two-electron bonds with an enhanced electron density accumulated over the shortest Nb−Ge13 contact in 1 . Low occupancies of the direct two-center two-electron (2c–2e) Nb−Ge and Ge−Ge σ bonds point to a considerable degree of electron delocalization over the Ge cages revealing their electron deficiency.     

大一学生实验室安全教育实效研究*

为考查线上线下混合安全教育模式实施后的实际效果,对天津理工大学化工、环安、机械、材料学院2020级大一学生进行实验室安全知识测试。结果表明:接受线上线下混合安全教育的大一学生安全知识的掌握情况较好;不足之处在于学生普遍对应急处置知识掌握较差。因此,在后续安全教育过程中,仍需进一步优化安全教育内容与模式,全面推进全员、全程安全教育。     

Phosphoric-Acid Retention in High-Temperature Proton-Exchange Membranes

Great efforts have been conducted to develop high temperature proton exchange membrane fuel cell (HT-PEMFC) due to its features of enhanced electrocatalyst reactivity, simplified hydrothermal management system and high CO tolerance of catalysts, and remarkable progress has been achieved. However, the easy leaching of phosphoric acid (PA) from the membranes during operation limits its commercial scale-up in complicated environments. This concept here mainly focuses on the recent developments for mitigation of PA loss in PEMs. The probable mechanisms of PA loss are proposed. The approaches to improve PA retention for example via introduction of phosphonic acid by covalent bond, using ion-pairs interaction and siphoning effect, and blending with inorganic nanoparticles are described in detail. Among these strategies, the siphoning effect from the intrinsic microporous PEMs is the most efficient and enables the cell to operate flexibly within a broad temperature range. Therefore, this concept may provide new ideas for the scientists to retain PA, to improve the cell performance and expand the potential applications of PA doped PEMs at elevated humidity and wide temperature range.     

基于快速神经网络架构搜索的鲁棒图像水印网络算法

为解决深度学习在图像水印算法中计算量大且模型冗余的问题，提高图像水印算法在抵抗噪声、旋转和剪裁等攻击时的鲁棒性，提出基于快速神经网络架构搜索（neural architecture search，NAS）的鲁棒图像水印网络算法。通过多项式分布学习快速神经网络架构搜索算法，在预设的搜索空间中搜索最优网络结构，进行图像水印的高效嵌入与鲁棒提取。首先，将子网络中线性连接的全卷积层设置为独立的神经单元结构，并参数化表示结构单元内节点的连接，预先设定结构单元内每个神经元操作的搜索空间；其次，在完成一个批次的数据集训练后，依据神经元操作中的被采样次数和平均损失函数值动态更新概率；最后，重新训练搜索完成的网络。水印网络模型的参数量较原始网络模型缩减了92%以上，大大缩短了模型训练时间。由于搜索得到的网络结构更为紧凑，本文算法具有较高的时间性能和较好的实验效果，在隐藏图像时，对空域信息的依赖比原始网络更少。对改进前后的2个网络进行了大量鲁棒性实验，对比发现，本文算法在CIFAR-10数据集上对抵抗椒盐噪声和旋转、移除像素行（列）等攻击优势显著；在ImageNet数据集上对抵抗椒盐高斯噪声、旋转、中值滤波、高斯滤波、JPEG压缩、裁剪等攻击优势显著，特别是对随机移除行（列）和椒盐噪声有较强的鲁棒性。     

Ultrahigh Loading Copper Single Atom Catalyst for Palladium-free Wacker Oxidation

Wacker oxidation is an industry-adopted process to transform olefins into value-added epoxides and carbonyls. However, traditional Wacker oxidation involves the use of homogeneous palladium and copper catalysts for the olefin addition and reductive elimination. Here, we demonstrated an ultrahigh loading Cu single atom catalyst(14% Cu, mass fraction) for the palladium-free Wacker oxidation of 4-vinylanisole into the corresponding ketone with N-methylhydroxylamine hydrochloride as an additive under mild conditions. Mechanistic studies by ¹⁸O and deuterium isotope labelling revealed a hydrogen shift mechanism in this palladium-free process using N-methylhydroxylamine hydrochloride as the oxygen source. The reaction scope can be further extended to Kucherov oxidation. Our study paves the way to replace noble metal catalysts in the traditional homogeneous processes with single atom catalysts.