【摘要】 提出了Siamese-BIGRU-Attention语义相似度计算方法，该方法在Siamese Network的基础框架上引入了BIGRU和attention模块，增强了前后文的捕捉能力并充分利用上下文信息挖掘因果关系，实现对输入文本更好的表征.实验表明：基于Siamese-BIGRU-Attention的语义相似度计算方法展现了更好的性能，其准确率及F1 score值均得到了一定的提升.更多还原

【Abstract】 The Siamese-BIGRU-Attention semantic similarity calculation method is proposed. It introduces BIGRU and Attention modules on the basis of Siamese Network, enhances the ability to capture contextual information and makes full use of contextual information to mining causal relationship to realize better representation of input text. Experiment results show that the semantic similarity calculation method based on Siamese-Bigru-attention presents better performance with significantly improved accuracy F1 score.更多还原