ID 原文 译文
56738 本文首先分析了嵌入方法在单网络节点表示学习中的研究现状,对比了现有方法的优劣. Providing thatnetwork data have changed in terms of their scale and modality, the research focus gradually shifted from singlenetwork mining to coupling network mining.
56739 然后借鉴单网络中嵌入方法的思想,针对耦合网络提出了耦合网络嵌入模型CWCNE. This paper first analyzes the research status of embedding methodsfor single networks and then compares their advantages and disadvantages.
56740 针对耦合网络的特性,改进了嵌入方法中的游走算法,提出了一种网络间带约束的随机游走策略; Furthermore, the paper presentsa model called CWCNE for coupling network embedding.
56741 同时改进了模型的训练方法,使用网络间迭代训练的方式来学习模型参数. The random walk and training algorithms of theproposed model are improved to adapt to coupling network features.
56742 最后使用了社交耦合网络、学术耦合网络、影视耦合网络、诗词耦合网络、著作耦合网络等5组数据集验证了CWCNE的有效性. The validity of the proposed model wasverified using social, academic, film, poetry, and work coupling network data.
56743 并在社团划分、实体识别、标签分类等任务上取得了良好的结果. Good results were obtained oncommunity detection, entity recognition, and label classification tasks.
56744 在多任务学习问题中,随机效应(random effects)可能同时存在于所有子任务中,而每个子任务又存在对应的稀疏效应(sparse effects). In multi-task learning scenarios, random effects may be shared among different tasks while each taskcan have its own sparse effects.
56745 这在文本分析尤其在对电影评论的情感分析中,尤为常见. This structure has often been observed in the field of sentiment analysis for movierating.
56746 在本文中,我们提出一种用于数据中同时存在共享随机效应和特定稀疏效应的混合多任务学习模型,并命名为MSS (multi-task learning with shared random effects and specific sparse effects)模型.在模型的建立过程中,我们利用Bayes框架,针对不同效应的特点设定不同的先验分布和超参数. In this study, we consider a multi-task learning problem in the presence of variables with shared randomeffects and specific sparse effects. To address this issue, we propose MSS (multi-task learning with shared randomeffects and specific sparse effects).
56747 在模型的求解过程中,我们使用变分推断克服Bayes推断中的计算难题,使MSS模型在大规模数据分析中具备广泛的适应性. To build this model, appropriate priors for the shared effects and specific effectsunder the Bayesian framework are considered.