Tabular data: Deep learning is not all you need

A key element in solving real-life data science problems is selecting the types of models to use. Tree ensemble models (such as XGBoost) are usually recommended for classification and regression problems with tabular data. However, several deep learning models for tabular data have recently been proposed, claiming to outperform XGBoost for some use cases. This paper explores whether these deep models should be a recommended option for tabular data by rigorously comparing the new deep models to XGBoost on various datasets. In addition to systematically comparing their performance, we consider the tuning and computation they require. Our study shows that XGBoost outperforms these deep models across the datasets, including the datasets used in the papers that proposed the deep models. We also demonstrate that XGBoost requires much less tuning. On the positive side, we show that an ensemble of deep models and XGBoost performs better on these datasets than XGBoost alone.     

基于卷积神经网络的5G蜂窝网络无线定位方法

5G蜂窝网络发展迅猛，其覆盖面积将逐渐增大，因此使用5G蜂窝网络进行定位是有研究潜力的研究方向。本文提出一种新的深度学习技术来实现高效、高精度和低占用的定位，以代替传统指纹定位过程中繁重的指纹库生成以及距离计算。该方法建立了一个特殊的卷积神经网络，并根据5G天线信号的接收信号强度指示、相位和到达角等特征量，选择合适的输入数据格式构造样本组建训练集，对该卷积神经网络进行训练。训练得到的卷积神经网络可以替代指纹定位中的庞大指纹库，非常有利于直接在5G移动设备端实现定位。虽然卷积神经网络在训练过程中需要大量时间，但在训练完毕后直接进行分类定位的速度非常快，可以保障定位实现的实时性。本文所实现的卷积神经网络权重与偏置所占内存不到0.5 MB，且能够在实际应用环境中以95%的定位准确率以及0.1 m的平均定位精度实现高精度定位。     

疫情下电磁场理论课程的网络教学

本文主要总结了新冠疫情期间作者的电磁场理论课程在线教学经验。对比分析了录播和直播的优缺点后,选择录播教学方式。基于超星网络教学平台,展示了录播网络教学的具体措施,包括网上答疑和学习效果检查以及在线批改作业等。给出了网络教学可以为线下教学继续使用的方法和手段,为疫情结束后的正常教学提供了新的网络教学补充措施。     

5G核心网业务模型的智能化预测研究

摘 要：核心网业务模型的建立是5G网络容量规划和网络建设的基础，通过现有方法得到的理论业务模型是静态不可变的且与实际网络存在偏离。为了克服现有5G核心网业务模型与现网模型适配性较差以及规划设备无法满足用户实际业务需求的问题，提出了一种长短期记忆（long short-term memory，LSTM）网络与卷积LSTM （convolution LSTM，ConvLSTM）网络双通道融合的 5G 核心网业务模型预测方法。该方法基于人工智能（artificial intelligence，AI）技术以实现高质量的核心网业务模型的智能预测，形成数据反馈闭环，实现网络自优化调整，助力网络智能化建设。     

基于VMD和BSA-KELM的高陡边坡位移预测模型研究

边坡位移的时间序列曲线存在复杂的非线性特性,传统的预测模型精度不足以满足预测要求。为此提出了基于变分模态分解的鸟群优化-核极限学习机的预测模型,并用于河北省某水泥厂的边坡位移预测。该方法首先采用VMD把边坡位移序列分解为一系列的有限带宽的子序列,再对各子序列分别采用相空间重构并用核极限学习机预测,采用鸟群算法优化相空间重构的嵌入维度和KELM中惩罚系数和核参数三个数值,以取得最优预测模型。最后将各个子序列预测值叠加,得到边坡位移的最终预测值。结果表明：和KELM、BSA-KELM、EEMD-BSA-KELM模型相比,基于VMD的BSA-KELM预测精度更高,为边坡位移的预测提供一种有效的方法。     

Magnetic properties,phase evolution,hollow structure and biomedical application of hematite (α-Fe2O3) and QUAIPH

We investigate synthesis, phase evolution, hollow and porous structure and magnetic properties of quasi-amorphous intermediate phase (QUAIPH) and hematite (α-Fe₂O₃) nanostructure synthesized by annealing of akaganeite (β-FeOOH) nanorods. It is found that the annealing temperature determines the phase composition of the products, the crystal structure/size dictates the magnetic properties whereas the final nanorod morphology is determined by the starting material. Annealing of β-FeOOH at ～300 °C resulted in the formation of hollow QUAIPH nanorods. The synthesized material shows low-cytotoxicity, superparamagnetism and good transverse relaxivity, which is rarely reported for QUAIPH. The QUAIPH nanorods started to transform to porous hematite nanostructures at ～350 °C and phase transformation was completed at 600 °C. During the annealing, the crystal structure changed from monoclinic (akaganeite) to quasi-amorphous and rhombohedral (hematite). Unusually, the crystallite size first decreased (akaganeite → QUAIPH) and then increased (QUAIPH → hematite) during annealing whereas the nanorods retained particle shape. The magnetic properties of the samples changed from antiferromagnetic (akaganeite) to superparamagnetic with blocking temperature T_B = 84 K (QUAIPH) and finally to weak-ferromagnetic with the Morin transition at T_M = 244 K and high coercivity H_C = 1652 Oe (hematite). The low-cytotoxicity and MRI relaxivity (r₂ = 5.80 mM^?1 s^?1 (akaganeite), r₂ = 4.31 mM^?1 s^?1 (QUAIPH) and r₂ = 5.17 mM^?1 s^?1 (hematite)) reveal potential for biomedical applications.     

Experimental investigation on the invert stability of operating railway tunnels with different drainage systems using 3D printing technology

In recent years, the invert anomalies of operating railway tunnels in water-rich areas occur frequently, which greatly affect the transportation capacity of the railway lines. Tunnel drainage system is a crucial factor to ensure the invert stability by regulating the external water pressure (EWP). By means of a three-dimensional (3D) printing model, this paper experimentally investigates the deformation behavior of the invert for the tunnels with the traditional drainage system (TDS) widely used in China and its optimized drainage system (ODS) with bottom drainage function. Six test groups with a total of 110 test conditions were designed to consider the design factors and environmental factors in engineering practice, including layout of the drainage system, blockage of the drainage system and groundwater level fluctuation. It was found that there are significant differences in the water discharge, EWP and invert stability for the tunnels with the two drainage systems. Even with a dense arrangement of the external blind tubes, TDS was still difficult to eliminate the excessive EWP below the invert, which is the main cause for the invert instability. Blockage of drainage system further increased the invert uplift and aggravated the track irregularity, especially when the blockage degree is more than 50%. However, ODS can prevent these invert anomalies by reasonably controlling the EWP at tunnel bottom. Even when the groundwater level reached 60 m and the blind tubes were fully blocked, the invert stability can still be maintained and the railway track experienced a settlement of only 1.8 mm. Meanwhile, the on-site monitoring under several rainstorms further showed that the average EWP of the invert was controlled within 84 kPa, while the maximum settlement of the track slab was only 0.92 mm, which also was in good agreement with the results of model test.     

Machine fault diagnosis with small sample based on variational information constrained generative adversarial network

In actual engineering scenarios, limited fault data leads to insufficient model training and over-fitting, which negatively affects the diagnostic performance of intelligent diagnostic models. To solve the problem, this paper proposes a variational information constrained generative adversarial network (VICGAN) for effective machine fault diagnosis. Firstly, by incorporating the encoder into the discriminator to map the deep features, an improved generative adversarial network with stronger data synthesis capability is established. Secondly, to promote the stable training of the model and guarantee better convergence, a variational information constraint technique is utilized, which constrains the input signals and deep features of the discriminator using the information bottleneck method. In addition, a representation matching module is added to impose restrictions on the generator, avoiding the mode collapse problem and boosting the sample diversity. Two rolling bearing datasets are utilized to verify the effectiveness and stability of the presented network, which demonstrates that the presented network has an admirable ability in processing fault diagnosis with few samples, and performs better than state-of-the-art approaches.     

使能未来工厂的5G能力综述

5G系统将移动通信服务从移动电话、移动宽带和大规模机器通信扩展到新的应用领域，即所谓对通信服务有特殊要求的垂直领域。对使能未来工厂的5G能力进行了全面的分析总结，包括弹性网络架构、灵活频谱、超可靠低时延通信、时间敏感网络、安全和定位，而弹性网络架构又包括对网络切片、非公共网络、5G局域网和边缘计算的支持。希望从广度到深度，对相关的理论及技术应用做透彻、全面的梳理，对其挑战做清晰的总结，从而为相关研究和工程技术人员提供借鉴。