高光谱成像技术的飞速发展给非侵入式医学成像带来新的契机,但高光谱医学图像具有高维度、高冗余以及“图谱合一”的特点,亟需针对上述特点设计智能诊断算法。近年来,Transformer已经在高光谱医学图像处理领域得到广泛应用。然而,不同...高光谱成像技术的飞速发展给非侵入式医学成像带来新的契机,但高光谱医学图像具有高维度、高冗余以及“图谱合一”的特点,亟需针对上述特点设计智能诊断算法。近年来,Transformer已经在高光谱医学图像处理领域得到广泛应用。然而,不同仪器设备、不同采集操作所获得的高光谱医学图像差异较大,这给现有Transformer诊断模型的实际应用带来了巨大挑战。针对上述问题,本文提出了一种空-谱自注意力Transformer(S3AT),自适应挖掘像素与像素间、波段与波段间的内蕴联系,并在分类阶段融合多个视野下的预测结果。首先,在Transformer编码器中,设计一种空-谱自注意力机制,获取不同视野下高光谱图像上的关键空间信息和重要波段,并将不同视野下所获得的空-谱自注意力进行融合。其次,在模型分类阶段,将不同视野下的预测结果根据可学习权重进行加权融合,对图像进行综合预测。在In-vivo Human Brain和BloodCell HSI两个数据集上,本文算法总体分类精度分别达到82.25%和91.74%。实验结果表明,所提出的算法有效改善高光谱医学图像分类性能。展开更多
负荷预测是综合能源系统(integrated energy system,IES)能量管理和优化调度的基础,其预测精度直接关系到系统的整体运行性能。提出了一种基于Transformer网络和多任务学习的园区综合能源系统电-热短期负荷预测模型。首先对Transformer...负荷预测是综合能源系统(integrated energy system,IES)能量管理和优化调度的基础,其预测精度直接关系到系统的整体运行性能。提出了一种基于Transformer网络和多任务学习的园区综合能源系统电-热短期负荷预测模型。首先对Transformer网络和多任务学习结构的基本原理进行了介绍;然后通过基于随机森林的特征选择步骤提取反映负荷特性和变化规律的典型指标,构建多任务学习输入特征,基于Transformer网络构建多任务学习权值共享层,并通过全连接层输出多能负荷的预测值;最后通过实际园区微能源系统的数据验证所提方法和算法的有效性,结果表明本文所提模型可以充分学习电-热耦合特征,提高负荷预测的精度。展开更多
文章研究了基于Transformer模型的中文文本生成方法,重点探讨了Transformer模型的编码器-解码器结构及其工作原理。在详细分析了编码器和解码器的工作机制后,文章利用Hugging Face Transformers开源模型进行了中文文本生成实验。结果表...文章研究了基于Transformer模型的中文文本生成方法,重点探讨了Transformer模型的编码器-解码器结构及其工作原理。在详细分析了编码器和解码器的工作机制后,文章利用Hugging Face Transformers开源模型进行了中文文本生成实验。结果表明,该方法在自制数据集上取得了良好的效果,其准确率、精确率和召回率分别达到92.5%、91.8%和90.6%。该研究不仅拓展了中文自然语言处理的理论基础,还为实际应用提供了高效的技术支持。展开更多
To utilize CFBC Al-rich fly ash, a mild hydrochemical extraction process was investigated for recovery of alumina. An alumina extraction efficiency of 92.31%was attained using a 45%NaOH solution, an original caustic r...To utilize CFBC Al-rich fly ash, a mild hydrochemical extraction process was investigated for recovery of alumina. An alumina extraction efficiency of 92.31%was attained using a 45%NaOH solution, an original caustic ratio （molar ratio of Na2O to Al2O3 in the sodium aluminate solution） of 25, a molar ratio of CaO to SiO2 in the fly ash of 1.1, a liquid volume to solid mass ratio of 9, a reaction temperature of 280 ℃, and a residence time of 1 h when treating fly ash with an alumina to silica mass ratio （A/S） of 0.78 and an alumina content of 32.43%. Additionally, the alumina leaching mechanism was explored via structural and chemical analysis, which revealed that after alkaline digestion, the main solid phase containing silica was NaCaHSiO4 with a theoretical A/S of zero.展开更多
In recent years, Empirical mode decomposition and Hilbert spectral analysis have been combined to identify system parameters. Singular-Value Decomposition is pro- posed as a signal preprocessing technique of Hilbert-H...In recent years, Empirical mode decomposition and Hilbert spectral analysis have been combined to identify system parameters. Singular-Value Decomposition is pro- posed as a signal preprocessing technique of Hilbert-Huang Transform to extract modal parameters for closely spaced modes and low-energy components. The proposed method is applied to a simulated airplane model built in Automatic Dynamic Analysis of Mechanical Systems software. The results demonstrate that the identified modal parameters are in good agreement with the baseline model.展开更多
An effective near-field - far-field (NF - FF) transformation with spherical scanning for quasi-planar antennas from irregularly spaced data is developed in this paper. Two efficient approaches for evaluating the regul...An effective near-field - far-field (NF - FF) transformation with spherical scanning for quasi-planar antennas from irregularly spaced data is developed in this paper. Two efficient approaches for evaluating the regularly spaced spherical samples from the nonuniformly distributed ones are proposed and numerically compared. Both the approaches rely on a nonredundant sampling representation of the voltage measured by the probe, based on an oblate ellipsoidal modelling of the antenna under test. The former employs the singular value decomposition method to reconstruct the NF data at the points fixed by the nonredundant sampling representation and can be applied when the irregularly acquired samples lie on nonuniform parallels. The latter is based on an iterative technique and can be used also when such a hypothesis does not hold, but requires the existence of a biunique correspondence between the uniform and nonuniform samples, associ- ating at each uniform sampling point the nearest irregular one. Once the regularly spaced spherical samples have been recovered, the NF data needed by a probe compensated NF - FF transformation with spherical scanning are efficiently evaluated by using an optimal sampling interpolation algorithm. It is so possible to accurately compensate known posi- tioning errors in the NF - FF transformation with spherical scanning for quasi-planar antennas. Some numerical tests assessing the accuracy and the robustness of the proposed approaches are reported.展开更多
This paper compares the estimated electromagnetic forces due to short circuits in power transformers using two computational models. The first model is based on approximate analytical expressions of the electromagneti...This paper compares the estimated electromagnetic forces due to short circuits in power transformers using two computational models. The first model is based on approximate analytical expressions of the electromagnetic lbrces as they have been compiled in earlier versions of the IEC standard 60076-5. The second model is based on a finite element model of the transformer using the software FEMM （Finite Element Method Magnetics）. The paper shows how valid the analytical model is for design purposes. Results have been obtained and compared from both models in a number of actual power distribution transformers. It is possible to conclude that the analytical formulation provides satisfactory results for the design of power transformers compared to detailed finite element models. A tool has been designed for this purpose and the main features of it will be described in the paper.展开更多
文摘高光谱成像技术的飞速发展给非侵入式医学成像带来新的契机,但高光谱医学图像具有高维度、高冗余以及“图谱合一”的特点,亟需针对上述特点设计智能诊断算法。近年来,Transformer已经在高光谱医学图像处理领域得到广泛应用。然而,不同仪器设备、不同采集操作所获得的高光谱医学图像差异较大,这给现有Transformer诊断模型的实际应用带来了巨大挑战。针对上述问题,本文提出了一种空-谱自注意力Transformer(S3AT),自适应挖掘像素与像素间、波段与波段间的内蕴联系,并在分类阶段融合多个视野下的预测结果。首先,在Transformer编码器中,设计一种空-谱自注意力机制,获取不同视野下高光谱图像上的关键空间信息和重要波段,并将不同视野下所获得的空-谱自注意力进行融合。其次,在模型分类阶段,将不同视野下的预测结果根据可学习权重进行加权融合,对图像进行综合预测。在In-vivo Human Brain和BloodCell HSI两个数据集上,本文算法总体分类精度分别达到82.25%和91.74%。实验结果表明,所提出的算法有效改善高光谱医学图像分类性能。
文摘负荷预测是综合能源系统(integrated energy system,IES)能量管理和优化调度的基础,其预测精度直接关系到系统的整体运行性能。提出了一种基于Transformer网络和多任务学习的园区综合能源系统电-热短期负荷预测模型。首先对Transformer网络和多任务学习结构的基本原理进行了介绍;然后通过基于随机森林的特征选择步骤提取反映负荷特性和变化规律的典型指标,构建多任务学习输入特征,基于Transformer网络构建多任务学习权值共享层,并通过全连接层输出多能负荷的预测值;最后通过实际园区微能源系统的数据验证所提方法和算法的有效性,结果表明本文所提模型可以充分学习电-热耦合特征,提高负荷预测的精度。
文摘文章研究了基于Transformer模型的中文文本生成方法,重点探讨了Transformer模型的编码器-解码器结构及其工作原理。在详细分析了编码器和解码器的工作机制后,文章利用Hugging Face Transformers开源模型进行了中文文本生成实验。结果表明,该方法在自制数据集上取得了良好的效果,其准确率、精确率和召回率分别达到92.5%、91.8%和90.6%。该研究不仅拓展了中文自然语言处理的理论基础,还为实际应用提供了高效的技术支持。
基金Project (2012BAF03B01) supported by the National Science and Technology Support Program of ChinaProject (2011AA060701) supported by the Hi-tech Research and Development Program of China
文摘To utilize CFBC Al-rich fly ash, a mild hydrochemical extraction process was investigated for recovery of alumina. An alumina extraction efficiency of 92.31%was attained using a 45%NaOH solution, an original caustic ratio （molar ratio of Na2O to Al2O3 in the sodium aluminate solution） of 25, a molar ratio of CaO to SiO2 in the fly ash of 1.1, a liquid volume to solid mass ratio of 9, a reaction temperature of 280 ℃, and a residence time of 1 h when treating fly ash with an alumina to silica mass ratio （A/S） of 0.78 and an alumina content of 32.43%. Additionally, the alumina leaching mechanism was explored via structural and chemical analysis, which revealed that after alkaline digestion, the main solid phase containing silica was NaCaHSiO4 with a theoretical A/S of zero.
文摘In recent years, Empirical mode decomposition and Hilbert spectral analysis have been combined to identify system parameters. Singular-Value Decomposition is pro- posed as a signal preprocessing technique of Hilbert-Huang Transform to extract modal parameters for closely spaced modes and low-energy components. The proposed method is applied to a simulated airplane model built in Automatic Dynamic Analysis of Mechanical Systems software. The results demonstrate that the identified modal parameters are in good agreement with the baseline model.
文摘An effective near-field - far-field (NF - FF) transformation with spherical scanning for quasi-planar antennas from irregularly spaced data is developed in this paper. Two efficient approaches for evaluating the regularly spaced spherical samples from the nonuniformly distributed ones are proposed and numerically compared. Both the approaches rely on a nonredundant sampling representation of the voltage measured by the probe, based on an oblate ellipsoidal modelling of the antenna under test. The former employs the singular value decomposition method to reconstruct the NF data at the points fixed by the nonredundant sampling representation and can be applied when the irregularly acquired samples lie on nonuniform parallels. The latter is based on an iterative technique and can be used also when such a hypothesis does not hold, but requires the existence of a biunique correspondence between the uniform and nonuniform samples, associ- ating at each uniform sampling point the nearest irregular one. Once the regularly spaced spherical samples have been recovered, the NF data needed by a probe compensated NF - FF transformation with spherical scanning are efficiently evaluated by using an optimal sampling interpolation algorithm. It is so possible to accurately compensate known posi- tioning errors in the NF - FF transformation with spherical scanning for quasi-planar antennas. Some numerical tests assessing the accuracy and the robustness of the proposed approaches are reported.
文摘This paper compares the estimated electromagnetic forces due to short circuits in power transformers using two computational models. The first model is based on approximate analytical expressions of the electromagnetic lbrces as they have been compiled in earlier versions of the IEC standard 60076-5. The second model is based on a finite element model of the transformer using the software FEMM （Finite Element Method Magnetics）. The paper shows how valid the analytical model is for design purposes. Results have been obtained and compared from both models in a number of actual power distribution transformers. It is possible to conclude that the analytical formulation provides satisfactory results for the design of power transformers compared to detailed finite element models. A tool has been designed for this purpose and the main features of it will be described in the paper.
