预训练模型(PTM)通过利用复杂的预训练目标和大量的模型参数,可以有效地获得无标记数据中的丰富知识。而在多模态中,PTM的发展还处于初期。根据具体模态的不同,将目前大多数的多模态PTM分为图像‒文本PTM和视频‒文本PTM;根据数据融合方...预训练模型(PTM)通过利用复杂的预训练目标和大量的模型参数,可以有效地获得无标记数据中的丰富知识。而在多模态中,PTM的发展还处于初期。根据具体模态的不同,将目前大多数的多模态PTM分为图像‒文本PTM和视频‒文本PTM;根据数据融合方式的不同,还可将多模态PTM分为单流模型和双流模型两类。首先,总结了常见的预训练任务和验证实验所使用的下游任务;接着,梳理了目前多模态预训练领域的常见模型,并用表格列出各个模型的下游任务以及模型的性能和实验数据比较;然后,介绍了M6(Multi-Modality to Multi-Modality Multitask Mega-transformer)模型、跨模态提示调优(CPT)模型、VideoBERT(Video Bidirectional Encoder Representations from Transformers)模型和AliceMind(Alibaba’s collection of encoder-decoders from Mind)模型在具体下游任务中的应用场景;最后,总结了多模态PTM相关工作面临的挑战以及未来可能的研究方向。展开更多
In this paper,we investigate the periodic traveling wave solutions problem for a single population model with advection and distributed delay.By the bifurcation analysis method,we can obtain periodic traveling wave so...In this paper,we investigate the periodic traveling wave solutions problem for a single population model with advection and distributed delay.By the bifurcation analysis method,we can obtain periodic traveling wave solutions for this model under the influence of advection term and distributed delay.The obtained results indicate that weak kernel and strong kernel can both deduce the existence of periodic traveling wave solutions.Finally,we apply the main results in this paper to Logistic model and Nicholson’s blowflies model.展开更多
For calculating the thermal storage time for an annular tube with phase change material （PCM）, a novel method is proposed. The method is suitable for either low-temperature PCM or high-temperature PCM whose initial ...For calculating the thermal storage time for an annular tube with phase change material （PCM）, a novel method is proposed. The method is suitable for either low-temperature PCM or high-temperature PCM whose initial temperature is near the melting point. The deviation fit is smaller than 8% when the time is below 2x104 s. Comparison between the predictions and the reported experimental data of thermal storage time at same conditions is investigated and good agreements have been got. Based on this method, the performance of the thermal storage unit and the role of natural convection are also investigated. Results show a linear relation between the maximum amount of stored heat and thermal storage time, and their ratio increases with the height of the thermal storage unit. As the thickness of the cavity increases, natural convection plays an increasingly important role in promoting the melting behavior of paraffin. When the thickness of the cavity is small, natural convection restrains the melting behavior of paraffin.展开更多
文摘预训练模型(PTM)通过利用复杂的预训练目标和大量的模型参数,可以有效地获得无标记数据中的丰富知识。而在多模态中,PTM的发展还处于初期。根据具体模态的不同,将目前大多数的多模态PTM分为图像‒文本PTM和视频‒文本PTM;根据数据融合方式的不同,还可将多模态PTM分为单流模型和双流模型两类。首先,总结了常见的预训练任务和验证实验所使用的下游任务;接着,梳理了目前多模态预训练领域的常见模型,并用表格列出各个模型的下游任务以及模型的性能和实验数据比较;然后,介绍了M6(Multi-Modality to Multi-Modality Multitask Mega-transformer)模型、跨模态提示调优(CPT)模型、VideoBERT(Video Bidirectional Encoder Representations from Transformers)模型和AliceMind(Alibaba’s collection of encoder-decoders from Mind)模型在具体下游任务中的应用场景;最后,总结了多模态PTM相关工作面临的挑战以及未来可能的研究方向。
基金Supported by the National Natural Science Foundation of China(12261050)Science and Technology Project of Department of Education of Jiangxi Province(GJJ2201612 and GJJ211027)Natural Science Foundation of Jiangxi Province of China(20212BAB202021)。
文摘In this paper,we investigate the periodic traveling wave solutions problem for a single population model with advection and distributed delay.By the bifurcation analysis method,we can obtain periodic traveling wave solutions for this model under the influence of advection term and distributed delay.The obtained results indicate that weak kernel and strong kernel can both deduce the existence of periodic traveling wave solutions.Finally,we apply the main results in this paper to Logistic model and Nicholson’s blowflies model.
基金Projects(51666006,51406071,51174105,51366005)supported by the National Natural Science Foundation of ChinaProject(2014CB460605)supported by the National Basic Research Program of China
文摘For calculating the thermal storage time for an annular tube with phase change material （PCM）, a novel method is proposed. The method is suitable for either low-temperature PCM or high-temperature PCM whose initial temperature is near the melting point. The deviation fit is smaller than 8% when the time is below 2x104 s. Comparison between the predictions and the reported experimental data of thermal storage time at same conditions is investigated and good agreements have been got. Based on this method, the performance of the thermal storage unit and the role of natural convection are also investigated. Results show a linear relation between the maximum amount of stored heat and thermal storage time, and their ratio increases with the height of the thermal storage unit. As the thickness of the cavity increases, natural convection plays an increasingly important role in promoting the melting behavior of paraffin. When the thickness of the cavity is small, natural convection restrains the melting behavior of paraffin.
