In recent years,semantic segmentation on 3D point cloud data has attracted much attention.Unlike 2D images where pixels distribute regularly in the image domain,3D point clouds in non-Euclidean space are irregular and...In recent years,semantic segmentation on 3D point cloud data has attracted much attention.Unlike 2D images where pixels distribute regularly in the image domain,3D point clouds in non-Euclidean space are irregular and inherently sparse.Therefore,it is very difficult to extract long-range contexts and effectively aggregate local features for semantic segmentation in 3D point cloud space.Most current methods either focus on local feature aggregation or long-range context dependency,but fail to directly establish a global-local feature extractor to complete the point cloud semantic segmentation tasks.In this paper,we propose a Transformer-based stratified graph convolutional network(SGT-Net),which enlarges the effective receptive field and builds direct long-range dependency.Specifically,we first propose a novel dense-sparse sampling strategy that provides dense local vertices and sparse long-distance vertices for subsequent graph convolutional network(GCN).Secondly,we propose a multi-key self-attention mechanism based on the Transformer to further weight augmentation for crucial neighboring relationships and enlarge the effective receptive field.In addition,to further improve the efficiency of the network,we propose a similarity measurement module to determine whether the neighborhood near the center point is effective.We demonstrate the validity and superiority of our method on the S3DIS and ShapeNet datasets.Through ablation experiments and segmentation visualization,we verify that the SGT model can improve the performance of the point cloud semantic segmentation.展开更多
To improve the computational efficieney of optimization based control methods, a new kind of Segmentized Optimization Strategy is presented,aiming at achieving more economical computation as well as comparatively sati...To improve the computational efficieney of optimization based control methods, a new kind of Segmentized Optimization Strategy is presented,aiming at achieving more economical computation as well as comparatively satisfactory performance. Its profitability is examined. And the effectiveaess is shown in the simulation.展开更多
Objective To investigate whether the very elderly patients with non-ST-segment elevation myocardial infarction (NSTEMI) will benefit from an invasive strategy versus a conservative strategy. Methods 190 consecutive pa...Objective To investigate whether the very elderly patients with non-ST-segment elevation myocardial infarction (NSTEMI) will benefit from an invasive strategy versus a conservative strategy. Methods 190 consecutive patients aged 80 years or older with NSTEMI were included in the retrospective study from September 2014 to August 2017, of which 69 patients received conservative strategy and 121 patients received invasive strategy. The primary outcome was death. Multivariate Cox regression models were used to assess the statistical association between strategies and mortality. The survival probability was further analyzed. Results The primary outcome occurred in 17.4% patients in the invasive group and in 42.0% patients in the conservative group (P = 0.0002). The readmission rate in the invasive group (14.9%) was higher than that in the conservative group (7.2%). Creatinine level (OR = 1.01, 95% CI: 0.10–1.03, P = 0.05) and use of diuretic (OR = 3.65, 95% CI: 1.56–8.53, P = 0.003) were independent influential factors for invasive strategy. HRs for multivariate Cox regression models were 3.45 (95% CI: 1.77–6.75, P = 0.0003), 3.02 (95% CI: 1.52–6.01, P = 0.0017), 2.93 (95% CI: 1. 46–5.86, P = 0.0024) and 2.47 (95% CI: 1.20–5.07, P = 0.0137). Compared with the patients received invasive strategy, the conservative group had remarkably reduced survival probability with time since treatment (P < 0.001). Conclusions An invasive strategy is superior to a conservative strategy in reducing mortality of patients aged 80 years or older with NSTEMI. Our results suggest that an invasive strategy is more suitable for the very elderly patients with NSTEMI in China.展开更多
ESA is an unsupervised approach to word segmentation previously proposed by Wang, which is an iterative process consisting of three phases: Evaluation, Selection and Adjustment. In this article, we propose Ex ESA, the...ESA is an unsupervised approach to word segmentation previously proposed by Wang, which is an iterative process consisting of three phases: Evaluation, Selection and Adjustment. In this article, we propose Ex ESA, the extension of ESA. In Ex ESA, the original approach is extended to a 2-pass process and the ratio of different word lengths is introduced as the third type of information combined with cohesion and separation. A maximum strategy is adopted to determine the best segmentation of a character sequence in the phrase of Selection. Besides, in Adjustment, Ex ESA re-evaluates separation information and individual information to overcome the overestimation frequencies. Additionally, a smoothing algorithm is applied to alleviate sparseness. The experiment results show that Ex ESA can further improve the performance and is time-saving by properly utilizing more information from un-annotated corpora. Moreover, the parameters of Ex ESA can be predicted by a set of empirical formulae or combined with the minimum description length principle.展开更多
The effective cathode flow field design can realize the internal water balance and higher current density output of proton exchange membrane fuel cells(PEMFC).Therefore,a segmented water management flow field is propo...The effective cathode flow field design can realize the internal water balance and higher current density output of proton exchange membrane fuel cells(PEMFC).Therefore,a segmented water management flow field is proposed in this study,i.e.a half separated-half coupled cathode(HSHC)flow field which has two inlets but just one outlet.A 3D numerical PEMFC model is applied to study the effect of the HSHC flow field on PEMFC performance and its operating strategy in terms of operating conditions.The study results are shown as follows:Compared with the two conventional cathode flow fields,the HSHC flow field improves the water balance along the channel and increases the current density by 17.1%at a cathode stoichiometry of 3.25.It is because the HSHC flow field can overcome the water loss at channels upstream and the water accumulation at channels downstream.The draw water phenomenon(DWP)in the HSHC flow field is observed,which is mainly affected by the water vapor pressure of channel.Based on the DWP,cooling channel inlet flow rates can be used to adjust water balance,but severe water loss should be avoided.In addition,the inlet temperature control in HSHC flow field should be that cell temperature>cathode channel inlet temperature>cooling channel inlet temperature>ambient temperatures for better water balance.展开更多
研究探讨了使用预训练的Pegasus模型进行长文本摘要时,不同文本分割方法对摘要质量的影响。收集来自知网的200篇关于STM32单片机的学术论文作为实验文本,比较了滑动窗口、句子分割、段落分割及滑动窗口加句子分割四种分割法的长文本摘...研究探讨了使用预训练的Pegasus模型进行长文本摘要时,不同文本分割方法对摘要质量的影响。收集来自知网的200篇关于STM32单片机的学术论文作为实验文本,比较了滑动窗口、句子分割、段落分割及滑动窗口加句子分割四种分割法的长文本摘要生成效果。实验使用ROUGE(Recall-Oriented Understudy for Gisting Evaluation)指标对生成的摘要进行评估,并对实验结果进行了详细分析。在生成摘要的质量方面,段落分割法表现出色,其ROUGE-1、ROUGE-2和ROUGE-L评分分别达到了30.85、7.60和20.15,轻微超过了句子分割法的评分,且显著优于句子分割加滑动窗口法。该研究旨在为研究者和开发者提供关于长文本摘要的实践经验和见解。展开更多
基金supported in part by the National Natural Science Foundation of China under Grant Nos.U20A20197,62306187the Foundation of Ministry of Industry and Information Technology TC220H05X-04.
文摘In recent years,semantic segmentation on 3D point cloud data has attracted much attention.Unlike 2D images where pixels distribute regularly in the image domain,3D point clouds in non-Euclidean space are irregular and inherently sparse.Therefore,it is very difficult to extract long-range contexts and effectively aggregate local features for semantic segmentation in 3D point cloud space.Most current methods either focus on local feature aggregation or long-range context dependency,but fail to directly establish a global-local feature extractor to complete the point cloud semantic segmentation tasks.In this paper,we propose a Transformer-based stratified graph convolutional network(SGT-Net),which enlarges the effective receptive field and builds direct long-range dependency.Specifically,we first propose a novel dense-sparse sampling strategy that provides dense local vertices and sparse long-distance vertices for subsequent graph convolutional network(GCN).Secondly,we propose a multi-key self-attention mechanism based on the Transformer to further weight augmentation for crucial neighboring relationships and enlarge the effective receptive field.In addition,to further improve the efficiency of the network,we propose a similarity measurement module to determine whether the neighborhood near the center point is effective.We demonstrate the validity and superiority of our method on the S3DIS and ShapeNet datasets.Through ablation experiments and segmentation visualization,we verify that the SGT model can improve the performance of the point cloud semantic segmentation.
文摘To improve the computational efficieney of optimization based control methods, a new kind of Segmentized Optimization Strategy is presented,aiming at achieving more economical computation as well as comparatively satisfactory performance. Its profitability is examined. And the effectiveaess is shown in the simulation.
文摘Objective To investigate whether the very elderly patients with non-ST-segment elevation myocardial infarction (NSTEMI) will benefit from an invasive strategy versus a conservative strategy. Methods 190 consecutive patients aged 80 years or older with NSTEMI were included in the retrospective study from September 2014 to August 2017, of which 69 patients received conservative strategy and 121 patients received invasive strategy. The primary outcome was death. Multivariate Cox regression models were used to assess the statistical association between strategies and mortality. The survival probability was further analyzed. Results The primary outcome occurred in 17.4% patients in the invasive group and in 42.0% patients in the conservative group (P = 0.0002). The readmission rate in the invasive group (14.9%) was higher than that in the conservative group (7.2%). Creatinine level (OR = 1.01, 95% CI: 0.10–1.03, P = 0.05) and use of diuretic (OR = 3.65, 95% CI: 1.56–8.53, P = 0.003) were independent influential factors for invasive strategy. HRs for multivariate Cox regression models were 3.45 (95% CI: 1.77–6.75, P = 0.0003), 3.02 (95% CI: 1.52–6.01, P = 0.0017), 2.93 (95% CI: 1. 46–5.86, P = 0.0024) and 2.47 (95% CI: 1.20–5.07, P = 0.0137). Compared with the patients received invasive strategy, the conservative group had remarkably reduced survival probability with time since treatment (P < 0.001). Conclusions An invasive strategy is superior to a conservative strategy in reducing mortality of patients aged 80 years or older with NSTEMI. Our results suggest that an invasive strategy is more suitable for the very elderly patients with NSTEMI in China.
基金supported in part by National Science Foundation of China under Grants No. 61303105 and 61402304the Humanity & Social Science general project of Ministry of Education under Grants No.14YJAZH046+2 种基金the Beijing Natural Science Foundation under Grants No. 4154065the Beijing Educational Committee Science and Technology Development Planned under Grants No.KM201410028017Beijing Key Disciplines of Computer Application Technology
文摘ESA is an unsupervised approach to word segmentation previously proposed by Wang, which is an iterative process consisting of three phases: Evaluation, Selection and Adjustment. In this article, we propose Ex ESA, the extension of ESA. In Ex ESA, the original approach is extended to a 2-pass process and the ratio of different word lengths is introduced as the third type of information combined with cohesion and separation. A maximum strategy is adopted to determine the best segmentation of a character sequence in the phrase of Selection. Besides, in Adjustment, Ex ESA re-evaluates separation information and individual information to overcome the overestimation frequencies. Additionally, a smoothing algorithm is applied to alleviate sparseness. The experiment results show that Ex ESA can further improve the performance and is time-saving by properly utilizing more information from un-annotated corpora. Moreover, the parameters of Ex ESA can be predicted by a set of empirical formulae or combined with the minimum description length principle.
基金the financial support for this research from the National Natural Science Foundation of China(52176063)the International Science and Technology projects of Huangpu District of Guangzhou City(2020GH08)the Guangzhou Science and Technology Plan Project(201907010036)。
文摘The effective cathode flow field design can realize the internal water balance and higher current density output of proton exchange membrane fuel cells(PEMFC).Therefore,a segmented water management flow field is proposed in this study,i.e.a half separated-half coupled cathode(HSHC)flow field which has two inlets but just one outlet.A 3D numerical PEMFC model is applied to study the effect of the HSHC flow field on PEMFC performance and its operating strategy in terms of operating conditions.The study results are shown as follows:Compared with the two conventional cathode flow fields,the HSHC flow field improves the water balance along the channel and increases the current density by 17.1%at a cathode stoichiometry of 3.25.It is because the HSHC flow field can overcome the water loss at channels upstream and the water accumulation at channels downstream.The draw water phenomenon(DWP)in the HSHC flow field is observed,which is mainly affected by the water vapor pressure of channel.Based on the DWP,cooling channel inlet flow rates can be used to adjust water balance,but severe water loss should be avoided.In addition,the inlet temperature control in HSHC flow field should be that cell temperature>cathode channel inlet temperature>cooling channel inlet temperature>ambient temperatures for better water balance.
文摘研究探讨了使用预训练的Pegasus模型进行长文本摘要时,不同文本分割方法对摘要质量的影响。收集来自知网的200篇关于STM32单片机的学术论文作为实验文本,比较了滑动窗口、句子分割、段落分割及滑动窗口加句子分割四种分割法的长文本摘要生成效果。实验使用ROUGE(Recall-Oriented Understudy for Gisting Evaluation)指标对生成的摘要进行评估,并对实验结果进行了详细分析。在生成摘要的质量方面,段落分割法表现出色,其ROUGE-1、ROUGE-2和ROUGE-L评分分别达到了30.85、7.60和20.15,轻微超过了句子分割法的评分,且显著优于句子分割加滑动窗口法。该研究旨在为研究者和开发者提供关于长文本摘要的实践经验和见解。
