Parkinsonian oscillations and their suppression by closed-loop deep brain stimulation based on fuzzy concept

This paper provides an adaptive closed-loop strategy for suppressing the pathological oscillations of the basal ganglia based on a variable universe fuzzy algorithm. The pathological basal ganglia oscillations in the theta (4-9 Hz) and beta (12-35 Hz) frequency bands have been demonstrated to be associated with the tremor and rigidity/bradykinesia symptoms in Parkinson's disease (PD). Although the clinical application of open-loop deep brain stimulation (DBS) is effective, the stimulation waveform with the fixed parameters cannot be self-adjusted as the disease progresses, and thus the stimulation effects go poor. To deal with this difficult problem, a variable universe fuzzy closed-loop strategy is proposed to modulate different PD states. We establish a cortico-basal ganglia-thalamocortical network model to simulate pathological oscillations and test the control effect. The results suggest that the proposed closed-loop control strategy can accommodate the variation of brain states and symptoms, which may become an alternative method to administrate the symptoms in PD.     

Fast prediction of aerodynamic noise induced by the flow around a cylinder based on deep neural network

Accurate and fast prediction of aerodynamic noise has always been a research hotspot in fluid mechanics and aeroacoustics. The conventional prediction methods based on numerical simulation often demand huge computational resources, which are difficult to balance between accuracy and efficiency. Here, we present a data-driven deep neural network (DNN) method to realize fast aerodynamic noise prediction while maintaining accuracy. The proposed deep learning method can predict the spatial distributions of aerodynamic noise information under different working conditions. Based on the large eddy simulation turbulence model and the Ffowcs Williams-Hawkings acoustic analogy theory, a dataset composed of 1216 samples is established. With reference to the deep learning method, a DNN framework is proposed to map the relationship between spatial coordinates, inlet velocity and overall sound pressure level. The root-mean-square-errors of prediction are below 0.82 dB in the test dataset, and the directivity of aerodynamic noise predicted by the DNN framework are basically consistent with the numerical simulation. This work paves a novel way for fast prediction of aerodynamic noise with high accuracy and has application potential in acoustic field prediction.     

Koopman analysis of nonlinear systems with a neural network representation

The observation and study of nonlinear dynamical systems has been gaining popularity over years in different fields. The intrinsic complexity of their dynamics defies many existing tools based on individual orbits, while the Koopman operator governs evolution of functions defined in phase space and is thus focused on ensembles of orbits, which provides an alternative approach to investigate global features of system dynamics prescribed by spectral properties of the operator. However, it is difficult to identify and represent the most relevant eigenfunctions in practice. Here, combined with the Koopman analysis, a neural network is designed to achieve the reconstruction and evolution of complex dynamical systems. By invoking the error minimization, a fundamental set of Koopman eigenfunctions are derived, which may reproduce the input dynamics through a nonlinear transformation provided by the neural network. The corresponding eigenvalues are also directly extracted by the specific evolutionary structure built in.     

4~20 K 温区固体界面导热填料接触热阻实验研究

超导量子干涉仪、 超导光子探测器等深空探测器需要液氦温区制冷技术提供极低温温度, 固体界面接触热阻的存在会增大耦合界面温度差, 进而增加制冷机系统冷损. 为定量探究4~20 K 深低温区固体接触热阻, 采用GM 作为冷源, 设计了一台可同时调节压力和低温温度的固体界面接触热阻测试实验台. 利用感压纸进行接触界面压力校核, 并对温度重复性进行验证. 实验测试了不同导热介质填充情况下, 温度和压力变化时固体接触热阻的变化规律. 基于最小二乘法对实验数据进行半经验公式拟合, 获得4 ~20 K 温区不同压力加载条件下的接触热阻的定量参考.     

基于NPU的实时深度学习跟踪算法实现北大核心CSCD

深度学习在检测领域高速发展,但受限于训练数据和计算效率,在基于嵌入式平台的边缘计算领域,尤其是实时跟踪应用中深度学习的智能化算法应用并不广泛。针对这一现象,同时为满足现阶段国产化、智能化的技术需求,提出了一种改进的孪生网络深度学习跟踪算法。在特征网络加入微调网络,解决了网络模型无法在线更新的问题,提升了跟踪的准确性;在IoUNet损失函数中加入中心距离惩罚项,解决了IoUNet当IoU相同时位置跳跃,存在收敛盲区和收敛速度慢的问题;将训练后的网络通过通道剪枝,缩减网络模型尺寸,提升了模型加载和运行的速度。在华为Atlas200NPU平台上实现了实时运行,算法准确率高达0.90(IoU>0.7),帧率达到66 Hz。     

Prediction model of K2CsSb photocathode reflectivity based on LSTM北大核心CSCD

针对目前K₂CsSb光阴极制备过程中无法预判光阴极生长状态的问题,提出一种基于长短期记忆（LSTM）循环神经网络的K₂CsSb光阴极反射率预测模型。一维原始反射率数据集经过清洗、筛选、序列化等预处理手段后重构为二维数据输入模型。为充分利用反射率数据在时序上高度相关的特性,采用双层LSTM网络提取特征,预测结果通过全连接层输出,以均方误差（MSE）作为模型预测效果的评判标准。实验结果表明,该模型的网络结构合理且在不同数据集下的表现良好,预测准确率可达99.21%。该模型可运用在K₂CsSb光阴极的制作过程中,通过反射率预测值反馈调节工艺参数以趋近目标走势,对提高光阴极性能具有促进作用。     

Composition and antioxidant activity of anthocyanins from Aronia melanocarpa extracted using an ultrasonic-microwave-assisted natural deep eutectic solvent extraction method

A time-saving, efficient, and environmentally friendly ultrasonic-microwave-assisted natural deep eutectic solvent (UMAE-NADES) extraction method was developed for the extraction of anthocyanins from Aronia melanocarpa. Eight different natural eutectic solvents were screened initially, and choline chloride-glycerol was selected as the extraction solvent. The extraction conditions were optimized using the response surface methodology, and the extraction rate of anthocyanins was higher than those achieved using the traditional ethanol method, natural deep eutectic solvent extraction method, and ultrasonic-microwave-assisted ethanol method. Six anthocyanins, including cyanidin-3-O-galactoside, cyanidin-3-O-glucoside, cyanidin-3-O-arabinoside, cyanidin-3-O-xyloside, cyanidin-3,5-O-dihexoside, and the dimer of cyanidin-hexoside were identified and extracted at a purity of 448.873 mg/g using high performance liquid chromatography-mass spectrometry (HPLC-MS). The compounds extracted using UMAE-NADES had higher antioxidant capacities than those extracted by the other three methods. The UMAE-NADES demonstrated significant efficiency toward the extraction of bioactive substances and has potential utility in the food and pharmaceutical industries.     

Development of an electronic stopping power model based on deep learning and its application in ion range prediction

Deep learning algorithm emerges as a new method to take the raw features from large dataset and mine their deep implicit relations, which is promising for solving traditional physical challenges. A particularly intricate and difficult challenge is the energy loss mechanism of energetic ions in solid, where accurate prediction of stopping power is a long-time problem. In this work, we develop a deep-learning-based stopping power model with high overall accuracy, and overcome the long-standing deficiency of the existing classical models by improving the predictive accuracy of stopping power for ultra-heavy ion with low energy, and the corresponding projected range. This electronic stopping power model, based on deep learning algorithm, could be hopefully applied for the study of ion-solid interaction mechanism and enormous relevant applications.     

Low-Temperature CH4 Total Oxidation on Catalysts Based on High Surface Area SnO2

Pure SnO₂, sulfated SnO₂-SO₄ ^2- and Pd supported on SnO₂ and SnO₂-SO₄ ^2- were prepared from SnO₂ precursor with high surface area, and used for CH₄ deep oxidation. The catalysts were characterized by means of N₂-BET, XRD, TG-DTA, XPS and TPD. SnO₂-SO₄ ^2- shows higher activity than SnO₂, due to the presence of more active oxygen species, superacid sites and its higher BET surface area. Pd/SnO₂ and Pd/SnO₂-SO₄ ^2- display essentially the same activity to each other, while it is much higher than the activity on SnO₂ and SnO₂-SO₄ ^2-. The main reason is ascribed to the concerted action between Pd and the supports.     

毛细管气相色谱法测定深海鱼油中的EPA和DHA含量

用氢氧化钾－甲醇酯化法将深海鱼油中的脂肪酸快速、有效地转化成脂肪酸甲酯,然后进行毛细管气相色谱法测定,以二十二碳饱和脂肪酸甲酯为内标物,建立了5,8,11,14,17－二十碳五烯酸（EPA）和4,7,10,13,16,19－二十二碳六烯酸（DHA）的定量分析方法,并用于实际样品分析。该方法对EPA和DHA定量分析的相对标准偏差分别为4．6％和5．0％,回收率分别为99．7％和99．4％。