seabet app download title: Nonlinear modeling and quantitative retrieval of large-scale multi-source heterogeneous data
seabet app download time:2020Year9month20日 Morning8:30-10:00
Reporting location: Electronic seabet app download of Railway College206
Reporter: Yang Zhan
seabet app download Summary:
In recent years，With the Internet，Internet of Things，Virtual reality and brain-computer interface and other technologies are developing rapidly，seabet app download is growing exponentially，Towards a better future for human societyProvide basic resources in the intelligent era of "human-machine-thing" integration，However，Heterogeneity、Multiple sources、Cross network、Inconsistency、Incompleteness has become todayMultiple sources across networksseabet app downloadMainFeatures，While enriching the description of the objective world, it also brings huge challenges to the field of seabet app download mining。Therefore，HowValid pairKnowledge contained in large-scale multi-source heterogeneous seabet app downloadProceedMiningsum search is the current seabet app download searchField ResearchOne of the important technologies in。This seabet app download mainly focuses on the difficulty of organizing large-scale heterogeneous data、Difficult to model、Difficult to integrate、Problems such as difficulty in retrieval，Proposed a novel nonlinear robust discrete hashing method。The main idea behind the model stems from the success of neural networks (i.e. nonlinear descriptors) in the field of representation learning，Use nonlinear descriptors instead of simple linear transformations，More consistent with the complex relationships between latent representations and heterogeneous multimedia data that are common in the real world。In the proposed model，We first learn a common latent representation via nonlinear descriptors，To encode complementarity and consistency information of heterogeneous multimedia data。The experimental results show the effectiveness of the proposed method。
