大规模科学装置与重大科学实验使得科学发现进入了数据密集型的第四范式,借助蓬勃发展的人工智能技术促进智能科学发现势在必行.机器学习作为人工智能中的一项重要技术,已广泛应用于各个科学领域.然而,现有工作仅研究特定任务下的机器...大规模科学装置与重大科学实验使得科学发现进入了数据密集型的第四范式,借助蓬勃发展的人工智能技术促进智能科学发现势在必行.机器学习作为人工智能中的一项重要技术,已广泛应用于各个科学领域.然而,现有工作仅研究特定任务下的机器学习方法,没能抽象出一个通用的智能科学发现研究框架.本文首先总结了科学发现任务中常用的机器学习方法,并将科学任务归类为五大机器学习问题.其次,提出了基于机器学习的智能科学发现研究框架,作为“AI for Science”的典型范例,阐述了一种高效的智能科学发现模式.再次,本文以时域天文学中发现瞬变事件这一科学任务为例,通过实验证明了唯有恰当地结合领域知识后,机器学习算法才能更好地服务于智能科学发现,验证了该框架的有效性.最后进行总结与展望,以期对各领域进行智能科学发现形成参考意义.展开更多
In order to explore the technology and effects of reducing nitrogen and potassium fertilizer applications in double-cropping rice,a field plot experiment was conducted to study the effects of optimized application of ...In order to explore the technology and effects of reducing nitrogen and potassium fertilizer applications in double-cropping rice,a field plot experiment was conducted to study the effects of optimized application of nitrogen and potassium fertilizers combined with returning Chinese milk vetch and straw to fields on yield,fertilizer utilization efficiency,net photosynthetic rate(Pn),stomatal conductance(Gs),intercellular CO_(2) concentration(Ci),chlorophyll content(SPAD value)and soil physical and chemical properties in late rice harvest period.The results showed that the optimized application of nitrogen and potassium fertilizers combined with the integrated technology of Chinese milk vetch and straw co-returning to the field could enhance the photosynthetic efficiency of double-cropping rice,increase rice yield,and enhance soil biological activity,especially T4 treatment involving the returning of Chinese milk vetch and straw to the field instead of 30%nitrogen fertilizer achieved the highest rice yield,fertilizer use efficiency,net photosynthetic rate and soil biological activity.Compared with the conventional fertilization treatment T2,the total rice yield of T4 treatment increased by 4.1%,among which the early rice and late rice increased by 6.3%and 2.4%,respectively;Pn,Gs and SPAD values of flag leaves at full heading stage significantly increased,and the contents of soil active organic carbon,alkali hydrolyzed nitrogen,available phosphorus and readily available potassium significantly increased.展开更多
文摘大规模科学装置与重大科学实验使得科学发现进入了数据密集型的第四范式,借助蓬勃发展的人工智能技术促进智能科学发现势在必行.机器学习作为人工智能中的一项重要技术,已广泛应用于各个科学领域.然而,现有工作仅研究特定任务下的机器学习方法,没能抽象出一个通用的智能科学发现研究框架.本文首先总结了科学发现任务中常用的机器学习方法,并将科学任务归类为五大机器学习问题.其次,提出了基于机器学习的智能科学发现研究框架,作为“AI for Science”的典型范例,阐述了一种高效的智能科学发现模式.再次,本文以时域天文学中发现瞬变事件这一科学任务为例,通过实验证明了唯有恰当地结合领域知识后,机器学习算法才能更好地服务于智能科学发现,验证了该框架的有效性.最后进行总结与展望,以期对各领域进行智能科学发现形成参考意义.
文摘In order to explore the technology and effects of reducing nitrogen and potassium fertilizer applications in double-cropping rice,a field plot experiment was conducted to study the effects of optimized application of nitrogen and potassium fertilizers combined with returning Chinese milk vetch and straw to fields on yield,fertilizer utilization efficiency,net photosynthetic rate(Pn),stomatal conductance(Gs),intercellular CO_(2) concentration(Ci),chlorophyll content(SPAD value)and soil physical and chemical properties in late rice harvest period.The results showed that the optimized application of nitrogen and potassium fertilizers combined with the integrated technology of Chinese milk vetch and straw co-returning to the field could enhance the photosynthetic efficiency of double-cropping rice,increase rice yield,and enhance soil biological activity,especially T4 treatment involving the returning of Chinese milk vetch and straw to the field instead of 30%nitrogen fertilizer achieved the highest rice yield,fertilizer use efficiency,net photosynthetic rate and soil biological activity.Compared with the conventional fertilization treatment T2,the total rice yield of T4 treatment increased by 4.1%,among which the early rice and late rice increased by 6.3%and 2.4%,respectively;Pn,Gs and SPAD values of flag leaves at full heading stage significantly increased,and the contents of soil active organic carbon,alkali hydrolyzed nitrogen,available phosphorus and readily available potassium significantly increased.
