自然语言生成(Natural Language Generation,NLG)任务是自然语言处理(Natural Languge Processing,NLP)任务中的一个子类,并且是一项具有挑战性的任务。随着深度学习在自然语言处理中的大量应用,其已经变成自然语言生成中处理各种任务...自然语言生成(Natural Language Generation,NLG)任务是自然语言处理(Natural Languge Processing,NLP)任务中的一个子类,并且是一项具有挑战性的任务。随着深度学习在自然语言处理中的大量应用,其已经变成自然语言生成中处理各种任务的主要方法。自然语言生成任务中主要有问答任务、生成摘要任务、生成评论任务、机器翻译任务、生成式对话任务等。传统的生成模型依赖输入文本,基于有限的知识生成文本。为解决这个问题,引入了知识增强的方法。首先介绍了自然语言生成的研究背景和重要模型,然后针对自然语言处理归纳介绍了提高模型性能的方法,以及基于内部知识(如提取关键词增强生成、围绕主题词等)和外部知识(如借助外部知识图谱增强生成)集成到文本生成过程中的方法和架构。最后,通过分析生成任务面临的一些问题,讨论了未来的挑战和研究方向。展开更多
The CuO nanocrystals were prepared by quick-precipitation and hydrolysis method respectively, using Cu(Ac)2 and NaOH as starting materials. The as-prepared CuO nanocrystals were characterized by XRD, TEM, XPS, UV-Visi...The CuO nanocrystals were prepared by quick-precipitation and hydrolysis method respectively, using Cu(Ac)2 and NaOH as starting materials. The as-prepared CuO nanocrystals were characterized by XRD, TEM, XPS, UV-Visible absorption spectroscopy and BET. The catalysis of CuO nanocrystals of different morphologies on ammonium perchlorate decomposition was investigated by thermal analysis. Results indicated the variation in temperature for NaOH addition played an important role in the shape of the CuO nanocrystals in precipitation method. Well-dispersed spherical CuO nanocrystals with an average size of 6 nm could be obtained when NaOH was added at 100 ℃ and spindle-shaped CuO nanocrystals exceeding 100 nm in diameter were obtained when NaOH was added at room temperature. Needle-shaped CuO nanocrystals could be prepared by the hydrolysis method, and the presence of small amounts of cetyltrimethylammonium bromide (CTAB) could make needle-shaped CuO well dispersed. The catalytic activity of CuO nanocrystals of different morphologies on ammonium perchlorate decomposition was high. The 2% of spherical CuO nanocrystals could make higher decomposition temperature of ammonium perchlorate decreased to 102 ℃. The exothermic quantity of decomposition was from 590 J·g-1 up to 1 420 J·g-1.展开更多
文摘自然语言生成(Natural Language Generation,NLG)任务是自然语言处理(Natural Languge Processing,NLP)任务中的一个子类,并且是一项具有挑战性的任务。随着深度学习在自然语言处理中的大量应用,其已经变成自然语言生成中处理各种任务的主要方法。自然语言生成任务中主要有问答任务、生成摘要任务、生成评论任务、机器翻译任务、生成式对话任务等。传统的生成模型依赖输入文本,基于有限的知识生成文本。为解决这个问题,引入了知识增强的方法。首先介绍了自然语言生成的研究背景和重要模型,然后针对自然语言处理归纳介绍了提高模型性能的方法,以及基于内部知识(如提取关键词增强生成、围绕主题词等)和外部知识(如借助外部知识图谱增强生成)集成到文本生成过程中的方法和架构。最后,通过分析生成任务面临的一些问题,讨论了未来的挑战和研究方向。
文摘The CuO nanocrystals were prepared by quick-precipitation and hydrolysis method respectively, using Cu(Ac)2 and NaOH as starting materials. The as-prepared CuO nanocrystals were characterized by XRD, TEM, XPS, UV-Visible absorption spectroscopy and BET. The catalysis of CuO nanocrystals of different morphologies on ammonium perchlorate decomposition was investigated by thermal analysis. Results indicated the variation in temperature for NaOH addition played an important role in the shape of the CuO nanocrystals in precipitation method. Well-dispersed spherical CuO nanocrystals with an average size of 6 nm could be obtained when NaOH was added at 100 ℃ and spindle-shaped CuO nanocrystals exceeding 100 nm in diameter were obtained when NaOH was added at room temperature. Needle-shaped CuO nanocrystals could be prepared by the hydrolysis method, and the presence of small amounts of cetyltrimethylammonium bromide (CTAB) could make needle-shaped CuO well dispersed. The catalytic activity of CuO nanocrystals of different morphologies on ammonium perchlorate decomposition was high. The 2% of spherical CuO nanocrystals could make higher decomposition temperature of ammonium perchlorate decreased to 102 ℃. The exothermic quantity of decomposition was from 590 J·g-1 up to 1 420 J·g-1.