In order to deal with the limitations during the register transfer level verification, a new functional verification method based on the random testing for the system-level of system-on-chip is proposed.The validity o...In order to deal with the limitations during the register transfer level verification, a new functional verification method based on the random testing for the system-level of system-on-chip is proposed.The validity of this method is proven theoretically.Specifically, testcases are generated according to many approaches of randomization.Moreover, the testbench for the system-level verification according to the proposed method is designed by using advanced modeling language.Therefore, under the circumstances that the testbench generates testcases quickly, the hardware/software co-simulation and co-verification can be implemented and the hardware/software partitioning planning can be evaluated easily.The comparison method is put to use in the evaluation approach of the testing validity.The evaluation result indicates that the efficiency of the partition testing is better than that of the random testing only when one or more subdomains are covered over with the area of errors, although the efficiency of the random testing is generally better than that of the partition testing.The experimental result indicates that this method has a good performance in the functional coverage and the cost of testing and can discover the functional errors as soon as possible.展开更多
This paper presents an analysis of the changes of the longitudinal and lateral profiles in the meander- ing reach of the Lower Wei River over the period from October 1973 to October 1976 during the course of degradati...This paper presents an analysis of the changes of the longitudinal and lateral profiles in the meander- ing reach of the Lower Wei River over the period from October 1973 to October 1976 during the course of degradation.Analysis results indicated that retrogressive erosion and subsequent downstream erosion occurred in the reach due to the lowering in the Tongguan elevation and the inflowing water carrying low sediment con- centrations.At the end of the degradation,the main channel widths of the majority ...展开更多
ReLM(Rephrasing Language Model)是当前性能领先的中文拼写纠错(CSC)模型。针对它在复杂语义场景中存在特征表达不足的问题,提出深层语义特征增强的ReLM——FeReLM(Feature-enhanced Rephrasing Language Model)。该模型利用深度可分...ReLM(Rephrasing Language Model)是当前性能领先的中文拼写纠错(CSC)模型。针对它在复杂语义场景中存在特征表达不足的问题,提出深层语义特征增强的ReLM——FeReLM(Feature-enhanced Rephrasing Language Model)。该模型利用深度可分离卷积(DSC)技术融合特征提取模型BGE(BAAI General Embeddings)生成的深层语义特征与ReLM生成的整体特征,从而有效提升模型对复杂上下文的解析力和拼写错误的识别纠正精度。首先,在Wang271K数据集上训练FeReLM,使模型持续学习句子中的深层语义和复杂表达;其次,迁移训练好的权重,从而将模型学习到的知识应用于新的数据集并进行微调。实验结果表明,在ECSpell和MCSC数据集上与ReLM、MCRSpell(Metric learning of Correct Representation for Chinese Spelling Correction)和RSpell(Retrieval-augmented Framework for Domain Adaptive Chinese Spelling Check)等模型相比,FeReLM的精确率、召回率、F1分数等关键指标的提升幅度可达0.6~28.7个百分点。此外,通过消融实验验证了所提方法的有效性。展开更多
亚中尺度过程是海洋学研究的前沿热点领域,从高分辨率资料中实现亚中尺度信号的快速提取对开展亚中尺度动力学研究具有重要意义。为此,根据亚中尺度过程的物理特性,提出一种基于深度学习的自动识别方法,构建了基于U-Net网络的海洋亚中...亚中尺度过程是海洋学研究的前沿热点领域,从高分辨率资料中实现亚中尺度信号的快速提取对开展亚中尺度动力学研究具有重要意义。为此,根据亚中尺度过程的物理特性,提出一种基于深度学习的自动识别方法,构建了基于U-Net网络的海洋亚中尺度过程识别网络(submesoscale processes automatic identification network, SM-Net),该网络采用视觉几何组网络作为主干特征提取网络并引入改进的混合注意力模块以提升识别能力。基于高分辨率MITgcm (Massachusetts Institute of Technology general circulation model)模式数据,通过SM-Net准确识别出南海东北部全年的亚中尺度过程,并分类为冷涡、暖涡和锋面。南海东北部亚中尺度冷涡、暖涡和锋面均多发生于冬季,夏季的发生频率较低,但吕宋海峡的亚中尺度过程全年均较为活跃。除吕宋海峡外,亚中尺度冷涡夏季多发生于台湾岛西南海域、吕宋岛西南海域和吕宋岛沿岸,冬季多发生于南海北部陆坡陆架区;亚中尺度暖涡夏季多发生于吕宋岛沿岸,冬季在南海北部陆坡陆架区较为活跃;亚中尺度锋面的时空特征与冷涡相似,但黑潮流经区域的发生频率更高。亚中尺度过程罗斯贝数和动能的时空特征与发生频率具有较好的一致性,暖涡的动能、罗斯贝数和直径均弱于冷涡。上述识别方法在南海的成功运用,为应用SWOT (surface water and ocean topography)卫星数据研究亚中尺度过程提供了一定参考。展开更多
基金supported by the National High Technology Research and Development Program of China (863 Program) (2002AA1Z1490)Specialized Research Fund for the Doctoral Program of Higher Education (20040486049)the University Cooperative Research Fund of Huawei Technology Co., Ltd
文摘In order to deal with the limitations during the register transfer level verification, a new functional verification method based on the random testing for the system-level of system-on-chip is proposed.The validity of this method is proven theoretically.Specifically, testcases are generated according to many approaches of randomization.Moreover, the testbench for the system-level verification according to the proposed method is designed by using advanced modeling language.Therefore, under the circumstances that the testbench generates testcases quickly, the hardware/software co-simulation and co-verification can be implemented and the hardware/software partitioning planning can be evaluated easily.The comparison method is put to use in the evaluation approach of the testing validity.The evaluation result indicates that the efficiency of the partition testing is better than that of the random testing only when one or more subdomains are covered over with the area of errors, although the efficiency of the random testing is generally better than that of the partition testing.The experimental result indicates that this method has a good performance in the functional coverage and the cost of testing and can discover the functional errors as soon as possible.
基金Supported by the Natural Science Foundation of China (50409002)by the Science Fund for Creative Research Groups of the Natural Science Foundation of China (50221903).
文摘This paper presents an analysis of the changes of the longitudinal and lateral profiles in the meander- ing reach of the Lower Wei River over the period from October 1973 to October 1976 during the course of degradation.Analysis results indicated that retrogressive erosion and subsequent downstream erosion occurred in the reach due to the lowering in the Tongguan elevation and the inflowing water carrying low sediment con- centrations.At the end of the degradation,the main channel widths of the majority ...
文摘ReLM(Rephrasing Language Model)是当前性能领先的中文拼写纠错(CSC)模型。针对它在复杂语义场景中存在特征表达不足的问题,提出深层语义特征增强的ReLM——FeReLM(Feature-enhanced Rephrasing Language Model)。该模型利用深度可分离卷积(DSC)技术融合特征提取模型BGE(BAAI General Embeddings)生成的深层语义特征与ReLM生成的整体特征,从而有效提升模型对复杂上下文的解析力和拼写错误的识别纠正精度。首先,在Wang271K数据集上训练FeReLM,使模型持续学习句子中的深层语义和复杂表达;其次,迁移训练好的权重,从而将模型学习到的知识应用于新的数据集并进行微调。实验结果表明,在ECSpell和MCSC数据集上与ReLM、MCRSpell(Metric learning of Correct Representation for Chinese Spelling Correction)和RSpell(Retrieval-augmented Framework for Domain Adaptive Chinese Spelling Check)等模型相比,FeReLM的精确率、召回率、F1分数等关键指标的提升幅度可达0.6~28.7个百分点。此外,通过消融实验验证了所提方法的有效性。
文摘亚中尺度过程是海洋学研究的前沿热点领域,从高分辨率资料中实现亚中尺度信号的快速提取对开展亚中尺度动力学研究具有重要意义。为此,根据亚中尺度过程的物理特性,提出一种基于深度学习的自动识别方法,构建了基于U-Net网络的海洋亚中尺度过程识别网络(submesoscale processes automatic identification network, SM-Net),该网络采用视觉几何组网络作为主干特征提取网络并引入改进的混合注意力模块以提升识别能力。基于高分辨率MITgcm (Massachusetts Institute of Technology general circulation model)模式数据,通过SM-Net准确识别出南海东北部全年的亚中尺度过程,并分类为冷涡、暖涡和锋面。南海东北部亚中尺度冷涡、暖涡和锋面均多发生于冬季,夏季的发生频率较低,但吕宋海峡的亚中尺度过程全年均较为活跃。除吕宋海峡外,亚中尺度冷涡夏季多发生于台湾岛西南海域、吕宋岛西南海域和吕宋岛沿岸,冬季多发生于南海北部陆坡陆架区;亚中尺度暖涡夏季多发生于吕宋岛沿岸,冬季在南海北部陆坡陆架区较为活跃;亚中尺度锋面的时空特征与冷涡相似,但黑潮流经区域的发生频率更高。亚中尺度过程罗斯贝数和动能的时空特征与发生频率具有较好的一致性,暖涡的动能、罗斯贝数和直径均弱于冷涡。上述识别方法在南海的成功运用,为应用SWOT (surface water and ocean topography)卫星数据研究亚中尺度过程提供了一定参考。
