针对爆破行业的从业人员对爆破英文文献阅读难度大,而又希望快速获取文献核心信息的需求。本文基于BART神经网络进行优化,对于爆破英文文献进行针对性提取出生成式摘要,并与传统的抽取式摘要进行对比,结果表明基于BART神经网络的摘要结...针对爆破行业的从业人员对爆破英文文献阅读难度大,而又希望快速获取文献核心信息的需求。本文基于BART神经网络进行优化,对于爆破英文文献进行针对性提取出生成式摘要,并与传统的抽取式摘要进行对比,结果表明基于BART神经网络的摘要结果在流畅度和信息完整性上更加优秀,对于爆破英文文献的信息内容提取有很大的帮助。Aiming at the needs of blasting industry practitioners who find it difficult to read English literature on blasting but want to quickly obtain the core information of the literature, this paper optimizes the BART neural network and extracts generative summaries for English literature on blasting. The summaries are compared with traditional extractive summaries. The results show that the summary results based on the BART neural network are more excellent in fluency and information completeness, which is of great help in extracting information content from English literature on blasting.展开更多
文摘针对爆破行业的从业人员对爆破英文文献阅读难度大,而又希望快速获取文献核心信息的需求。本文基于BART神经网络进行优化,对于爆破英文文献进行针对性提取出生成式摘要,并与传统的抽取式摘要进行对比,结果表明基于BART神经网络的摘要结果在流畅度和信息完整性上更加优秀,对于爆破英文文献的信息内容提取有很大的帮助。Aiming at the needs of blasting industry practitioners who find it difficult to read English literature on blasting but want to quickly obtain the core information of the literature, this paper optimizes the BART neural network and extracts generative summaries for English literature on blasting. The summaries are compared with traditional extractive summaries. The results show that the summary results based on the BART neural network are more excellent in fluency and information completeness, which is of great help in extracting information content from English literature on blasting.
