Shanghai has experienced the greatest land subsidence in China in the past sixty years and produced undesirable environmental impact. However, horizontal ground deformation has not been understood yet. Therefore groun...Shanghai has experienced the greatest land subsidence in China in the past sixty years and produced undesirable environmental impact. However, horizontal ground deformation has not been understood yet. Therefore ground deformation monitoring together with the analysis of its driving forces are critical for geo-hazards early-warning, city planning and sustainable urbanization in Shanghai. In this paper, two-dimensional ground deformation monitoring was performed in Shanghai with SBAS and MSBAS InSAR methods. Twenty-nine Multi-Look Fine 6 （MF6） Radarsat-2 SLC data acquired during 2011-2013 were used to derive vertical ground deformation. Meanwhile, six descending Multi-Look Fine 6 （MF6） and four ascending Multi-Look Fine 2 （MF2） spanning April to August, 2008, were used to derive vertical and horizontal ground deformation during the observation period. The results indicate that vertical and horizontal deformations in 2008 were not homogeneously distributed in different districts ranging from 0-2 cm/year. Vertical deformation rate during 2011-2013 were decreased to less than 1 cm/year in most district of Shanghai area. Activities from groundwater exploitation and rapid urbanization are responsible for most of the ground deformation in Shanghai. Thus, future ground deformation in vertical and horizontal directions should be warranted.展开更多
目的随着人脸图像合成技术的快速发展,基于深度学习的人脸伪造技术对社会信息安全的负面影响日益增长。然而,由于不同伪造方法生成的样本之间的数据分布存在较大差异,现有人脸伪造检测方法准确性不高,泛化性较差。为了解决上述问题,提...目的随着人脸图像合成技术的快速发展,基于深度学习的人脸伪造技术对社会信息安全的负面影响日益增长。然而,由于不同伪造方法生成的样本之间的数据分布存在较大差异,现有人脸伪造检测方法准确性不高,泛化性较差。为了解决上述问题,提出一种多元软混合样本驱动的图文对齐人脸伪造检测新方法,充分利用图像与文本的多模态信息对齐,捕捉微弱的人脸伪造痕迹。方法考虑到传统人脸伪造检测方法仅在单一模式的伪造图像上训练,难以应对复杂伪造模式,提出了一种多元软混合的数据增广方式(multivariate and soft blending augmentation,MSBA),促进网络同时捕捉多种伪造模式线索的能力,增强了网络模型对复杂和未知的伪造模式的检测能力。由于不同人脸伪造图像的伪造模式与伪造力度多种多样,导致网络模型真伪检测性能下降。本文基于MSBA方式设计了多元伪造力度估计(multivariate forgery intensity estimation,MFIE)模块,有效针对不同模式和力度的人脸伪造图像进行学习,引导图像编码器提取更加具有泛化性的特征,提高了整体网络框架的检测准确性。结果在域内实验中,与对比算法性能最好的相比,本文方法的准确率(accuracy,ACC)与AUC(area under the curve)指标分别提升3.32%和4.02%。在跨域实验中,本文方法与6种典型方法在5个数据集上进行了性能测试与比较,平均AUC指标提高3.27%。消融实验结果表明本文提出的MSBA方式和MFIE模块对于人脸伪造检测性能的提升均有较好的表现。结论本文面向人脸伪造检测任务设计的CLIP(contrastive language-image pre-training)网络框架大大提高了人脸伪造检测的准确性,提出的MSBA方式和MFIE模块均起到了较好的助力效果,取得了超越已有方法的性能表现。展开更多
Intermittent demand refers to the specific demand pattern with frequent periods of zero demand.It occurs in a variety of industries including industrial equipment,automotive and specialty chemicals.In some industries ...Intermittent demand refers to the specific demand pattern with frequent periods of zero demand.It occurs in a variety of industries including industrial equipment,automotive and specialty chemicals.In some industries or some sectors of industry,even majority of products are in intermittent demand pattern.Due to the usually small and highly variable demand sizes,accurate forecasting of intermittent demand has always been challenging.However,accurate forecasting of intermittent demand is critical to the effective inventory management.In this study we present a band new method-modified TSB method for the forecasting of intermittent demand.The proposed method is based on TSB method,and adopts similar strategy,which has been used in m SBA method to update demand interval and demand occurrence probability when current demand is zero.To evaluate the proposed method,16289 daily demand records from the M5 data set that are identified as intermittent demands according to two criteria,and an empirical data set consisting three years’monthly demand history of 1718 medicine products are used.The proposed m TSB method achieves the best results on MASE and RMASE among all comparison methods on the M5 data set.On the empirical data set,the study shows that m TSB attains an ME of 0.07,which is the best among six comparison methods.Additionally,on the MSE measurement,m TSB shows a similar result as SES,both of which outperform other methods.展开更多
基金supported by the China Science National Foundation (No. 41372353)
文摘Shanghai has experienced the greatest land subsidence in China in the past sixty years and produced undesirable environmental impact. However, horizontal ground deformation has not been understood yet. Therefore ground deformation monitoring together with the analysis of its driving forces are critical for geo-hazards early-warning, city planning and sustainable urbanization in Shanghai. In this paper, two-dimensional ground deformation monitoring was performed in Shanghai with SBAS and MSBAS InSAR methods. Twenty-nine Multi-Look Fine 6 （MF6） Radarsat-2 SLC data acquired during 2011-2013 were used to derive vertical ground deformation. Meanwhile, six descending Multi-Look Fine 6 （MF6） and four ascending Multi-Look Fine 2 （MF2） spanning April to August, 2008, were used to derive vertical and horizontal ground deformation during the observation period. The results indicate that vertical and horizontal deformations in 2008 were not homogeneously distributed in different districts ranging from 0-2 cm/year. Vertical deformation rate during 2011-2013 were decreased to less than 1 cm/year in most district of Shanghai area. Activities from groundwater exploitation and rapid urbanization are responsible for most of the ground deformation in Shanghai. Thus, future ground deformation in vertical and horizontal directions should be warranted.
文摘目的随着人脸图像合成技术的快速发展,基于深度学习的人脸伪造技术对社会信息安全的负面影响日益增长。然而,由于不同伪造方法生成的样本之间的数据分布存在较大差异,现有人脸伪造检测方法准确性不高,泛化性较差。为了解决上述问题,提出一种多元软混合样本驱动的图文对齐人脸伪造检测新方法,充分利用图像与文本的多模态信息对齐,捕捉微弱的人脸伪造痕迹。方法考虑到传统人脸伪造检测方法仅在单一模式的伪造图像上训练,难以应对复杂伪造模式,提出了一种多元软混合的数据增广方式(multivariate and soft blending augmentation,MSBA),促进网络同时捕捉多种伪造模式线索的能力,增强了网络模型对复杂和未知的伪造模式的检测能力。由于不同人脸伪造图像的伪造模式与伪造力度多种多样,导致网络模型真伪检测性能下降。本文基于MSBA方式设计了多元伪造力度估计(multivariate forgery intensity estimation,MFIE)模块,有效针对不同模式和力度的人脸伪造图像进行学习,引导图像编码器提取更加具有泛化性的特征,提高了整体网络框架的检测准确性。结果在域内实验中,与对比算法性能最好的相比,本文方法的准确率(accuracy,ACC)与AUC(area under the curve)指标分别提升3.32%和4.02%。在跨域实验中,本文方法与6种典型方法在5个数据集上进行了性能测试与比较,平均AUC指标提高3.27%。消融实验结果表明本文提出的MSBA方式和MFIE模块对于人脸伪造检测性能的提升均有较好的表现。结论本文面向人脸伪造检测任务设计的CLIP(contrastive language-image pre-training)网络框架大大提高了人脸伪造检测的准确性,提出的MSBA方式和MFIE模块均起到了较好的助力效果,取得了超越已有方法的性能表现。
基金supported in part by XJTLU laboratory for intelligent computation and financial technology through XJTLU Key Programme Special Fund(KSFeP-02 and KSF-E-21)
文摘Intermittent demand refers to the specific demand pattern with frequent periods of zero demand.It occurs in a variety of industries including industrial equipment,automotive and specialty chemicals.In some industries or some sectors of industry,even majority of products are in intermittent demand pattern.Due to the usually small and highly variable demand sizes,accurate forecasting of intermittent demand has always been challenging.However,accurate forecasting of intermittent demand is critical to the effective inventory management.In this study we present a band new method-modified TSB method for the forecasting of intermittent demand.The proposed method is based on TSB method,and adopts similar strategy,which has been used in m SBA method to update demand interval and demand occurrence probability when current demand is zero.To evaluate the proposed method,16289 daily demand records from the M5 data set that are identified as intermittent demands according to two criteria,and an empirical data set consisting three years’monthly demand history of 1718 medicine products are used.The proposed m TSB method achieves the best results on MASE and RMASE among all comparison methods on the M5 data set.On the empirical data set,the study shows that m TSB attains an ME of 0.07,which is the best among six comparison methods.Additionally,on the MSE measurement,m TSB shows a similar result as SES,both of which outperform other methods.
