来自多源感知设备所采集的多模态交通数据,由于探测设备、网络、数据传输等错误往往存在丢失.交通数据的缺失对交通网络智能规划、避免拥堵等会产生重大的负面影响.同时,来自于不同平台数据的编码方式、标识存在差异,很大程度上影响了...来自多源感知设备所采集的多模态交通数据,由于探测设备、网络、数据传输等错误往往存在丢失.交通数据的缺失对交通网络智能规划、避免拥堵等会产生重大的负面影响.同时,来自于不同平台数据的编码方式、标识存在差异,很大程度上影响了交通数据的利用.基于此,本文针对交通监控视频与车流量探测数据,结合张量理论,建立了用以描述多模态交通数据的张量模型,并提出了基于Tucker-Crossover的多模态数据补全算法(Tucker-Crossover based Multimodal Data Imputation Algorithm,TCM D-IA),用于多模态交通缺失数据的补全.该方法利用Tucker分解后不同阶的因子矩阵和核矩阵进行相关性融合,从而提高缺失值估计效果.在真实交通数据集上的实验表明,TCMD-IA的多模态交通缺失数据补全效果优于其他方法,且具有较好的鲁棒性.展开更多
HIV- 1 RT is an important target for the treatment of AIDS. There are two major classes of antiviral agents that inhibit HIV- 1 RT have been identified, nucleoside RT inhibitors （NRTIs） and non-nucleoside RT inhibit...HIV- 1 RT is an important target for the treatment of AIDS. There are two major classes of antiviral agents that inhibit HIV- 1 RT have been identified, nucleoside RT inhibitors （NRTIs） and non-nucleoside RT inhibitors （NNRTIs）. In this report, a noval class of non-nucleoside compound with potential RT inhibitory activity were found from the traditional Chinese medicines database （TCMD） using a combination of virtual screening, docking, molecular dynamic simulations, where results were ranked by scoring function of the docking tool. The result indicates that M4753 （a compound derived from TCMD） has not only the lowest bonding energy but also the best match in geometric conformation with the forthcoming NNRTIs. Accordingly M4753 might possibly become a promising lead compound of NNRTIs for AIDS therapy.展开更多
文摘来自多源感知设备所采集的多模态交通数据,由于探测设备、网络、数据传输等错误往往存在丢失.交通数据的缺失对交通网络智能规划、避免拥堵等会产生重大的负面影响.同时,来自于不同平台数据的编码方式、标识存在差异,很大程度上影响了交通数据的利用.基于此,本文针对交通监控视频与车流量探测数据,结合张量理论,建立了用以描述多模态交通数据的张量模型,并提出了基于Tucker-Crossover的多模态数据补全算法(Tucker-Crossover based Multimodal Data Imputation Algorithm,TCM D-IA),用于多模态交通缺失数据的补全.该方法利用Tucker分解后不同阶的因子矩阵和核矩阵进行相关性融合,从而提高缺失值估计效果.在真实交通数据集上的实验表明,TCMD-IA的多模态交通缺失数据补全效果优于其他方法,且具有较好的鲁棒性.
基金supported by the grants from Chinese National Science Foundation(No.30472166)the Tianjin Commission of Sciences and Technology under the Contract(No.06YFGZSH07000)
文摘HIV- 1 RT is an important target for the treatment of AIDS. There are two major classes of antiviral agents that inhibit HIV- 1 RT have been identified, nucleoside RT inhibitors （NRTIs） and non-nucleoside RT inhibitors （NNRTIs）. In this report, a noval class of non-nucleoside compound with potential RT inhibitory activity were found from the traditional Chinese medicines database （TCMD） using a combination of virtual screening, docking, molecular dynamic simulations, where results were ranked by scoring function of the docking tool. The result indicates that M4753 （a compound derived from TCMD） has not only the lowest bonding energy but also the best match in geometric conformation with the forthcoming NNRTIs. Accordingly M4753 might possibly become a promising lead compound of NNRTIs for AIDS therapy.
