Subjects: Computer Science >> Integration Theory of Computer Science submitted time 2018-05-02 Cooperative journals: 《计算机应用研究》
Abstract： In view of the problems of topic transfer and the existence of a large number of short text in the dialogue content in open domain dialogue systems, the traditional similarity-based processing method has many limitations. This paper proposeed an innovative method, which is based on the relevance of the sentences to determine whether the dialogue topic transfer, and compares the difference between the correlation-based and the similarity-based methods in revealing the relationship between sentences. Furthermore, this paper presents a correlation-based algorithm to calculate the correlation of words and apply it to segment topics of sentences, and this can address some challenges of topic transfer detection. Comparing with existing methods, the experimental results demonstrate the superior performance of the correlation-based method in this paper.
Subjects: Agriculture, Forestry,Livestock & Aquatic Products Science >> Other Disciplines of Agriculture, Forestry,Livestock & Aquatic Products Science submitted time 2023-05-15 Cooperative journals: 《智慧农业（中英文）》
Abstract： Pre-training refers to the process of training deep neural network parameters on a large corpus before a specific task model performs a particular task. This approach enables downstream tasks to fine-tune the pre-trained model parameters based on a small amount of labeled data, eliminating the need to train a new model from scratch. Currently, research on named entity recognition (NER) using pre-trained language model (PLM) only uses the last layer of the PLM to express output when facing challenges such as complex entity naming methods and fuzzy entity boundaries in the agricultural field. This approach ignores the rich information contained in the internal layers of the model themselves. To address these issues, a named entity recognition method based on progressive convolutional networks has been proposed. This method stores natural sentences and outputs representations of each layer obtained through PLM. The intermediate outputs of the pre-trained model are sequentially convolved to extract shallow feature information that may have been overlooked previously. Using the progressive convolutional network module proposed in this research, the adjacent two-layer representations are convolved from the first layer, and the fusion result continues to be convolved with the next layer, resulting in enhanced sentence embedding that includes the entire information dimension of the model layer. The method does not require the introduction of external information, which makes the sentence representation contain richer information. Research has shown that the sentence embedding output of the model layer near the input contains more fine-grained information, such as phrases and phrases, which can assist with NER problems in the agricultural field. Fully utilizing the computational power already used, the results obtained can enhance the representation embedding of sentences. Finally, the conditional random field (CRF) model was used to generate the global optimal sequence. On a constructed agricultural dataset containing four types of agricultural entities, the proposed method's comprehensive indicator F1 value increased by 3.61% points compared to the basic BERT (Bidirectional Encoder Representation from Transformers) model. On the open dataset MSRA, the F1 value also increased to 94.96%, indicating that the progressive convolutional network can enhance the model's ability to represent natural language and has advantages in NER tasks.
Subjects: Biology >> Botany >> Applied botany submitted time 2022-03-18 Cooperative journals: 《广西植物》
Abstract： Camellia kweichowensis is a rare 5-locular capsule-bearing representative species of the section Camellia with biological importance and economic value. This study explored the reasons for its endangered status and effective ways to conserve and utilize resources. In this study, by combining the sample investigation and actual measurement methods, the analysis of population age structure and survivability, we studied the quantitative characteristics of population structure dynamics and future development trend. The results were as follows: (1) The primary vegetation of C. kweichowensis had typical characteristics of subtropical humid evergreen broad-leaved forests. While the mixed evergreen and deciduous broad-leaved forest accounted for the main stand, the coniferous broad-leaved mixed forest of Pinus armandii + Betula luminifera + Liquidambar formosana + Nyssa sinensis + Camellia sp. + Eurya sp. + Schima sp. was common. (2) The growth population of C. kweichowensis dominated the growth structure. The points were mainly concentrated in the small and medium tree stages, and the sum of the proportion constituted 73.02% of the overall population. Deevey-Ⅱ type characteristic of the population survival curve was obvious, the life expectancy of C. kweichowensis was the maximum at the seedling stage. The change trend of mortality and the vanishing curves of the same plot were approximately the same. The quantitative dynamic analysis indicated that the three plots had abundant seedling pools, but were sensitive to the disturbance outside the habitat. The probability of surviving to Ⅶ age-class dropped significantly to 36.17%. On the whole, the population structure showed the characteristics of “growth in the early stage and decline in the late stage of Ⅳ age-class”. (3) The time-series forecast analysis indicated that after the growth period of 2-, 4-, 6-, 8-, and 10-diameter scales, the C. kweichowensis population dynamics tilted toward the succession of middle trees (Ⅴ-Ⅶ age-class), big trees (Ⅷ-Ⅹ age-class), and aged trees (Ⅺ-ⅩⅢ age-class). In conclusion, due to the limited habitat resources and space, intraspecific and interspecific competition intensifies, which reduces the survival rate of population seedlings and the number of young, and increased the mortality of adult plants caused by human predatory logging. As a result, the population renewal and diffusion are hindered, the distribution area is narrow, and the species tend to be endangered.
Subjects: Biology >> Botany >> Applied botany submitted time 2019-03-14 Cooperative journals: 《广西植物》
Abstract：采用浸提法和GC-MS检测了贵州石笔木种子内源有机化合物种类、相对含量，以及种皮和胚乳在不同溶剂、温度和浓度条件下的浸提物活性。结果显示：（1）种皮和胚乳中皆含有有机酸、烯、酯、醇、醛、酚等6类相对含量较高的有机化合物，其中种皮含有机酸7种、烯类1种、酯类5种、醇类3种、醛类3种、酚类1种，胚乳含有机酸6种、烯类1种、酯类1种、醇类1种、醛类1种、酚类1种；（2）种皮浸提物活性显著高于胚乳浸提物活性(P<0.05)，且其烯、醛、醇和酯类含量分别高出胚乳含量的8.78%、2.66%、2.15%和1.70%，可能是对种子萌发起主要作用的内源抑制物质；（3）不同条件下处理的浸提液均能显著抑制白菜种子发芽及幼苗生长，浸提液抑制物活性表现为醇溶剂大于水溶剂，并随着浸提液浓度的升高而增大、随着浸提温度的升高而增强，在初始温度为100 ℃时，浸提液抑制活性达到最大值。贵州石笔木种子的内源有机化合物在种子萌发过程中发挥着不同程度的抑制作用，探索其与种子萌发的作用机制，解决种子萌发育苗的关键技术及在农林业生产中应用，这在植物种子的生物学特性及萌发生理研究方面具有重要意义。
Subjects: Astronomy >> Astrophysical processes submitted time 2019-12-20 Cooperative journals: 《天文研究与技术》
Abstract：太阳活动区是各类太阳活动的主要能量来源，剧烈的太阳活动会直接影响人类的生存环境，因此准确地检测与跟踪太阳活动区对监控和预报空间天气非常重要。本文基于深度学习框架的YOLOv3-spp和DeepSort提出了一种太阳活动区检测和跟踪方法(Active Regions Detection and Tracking Method, ARDTM)，该方法较好地解决了传统图像处理方法易将一个太阳活动区误检测为多个，或者多个太阳活动区误检测为一个活动区的问题；及时捕获到新产生的太阳活动区和终止跟踪消失的太阳活动区，有效地提高了太阳活动区的跟踪准确率。实验结果表明该方法可以较好地检测和跟踪不同望远镜、不同时间间隔序列图像中的太阳活动区。
Subjects: Biology >> Botany >> Applied botany submitted time 2022-06-07 Cooperative journals: 《广西植物》
Abstract： Castor root rot is a root disease caused by Fusarium solani, which seriously threatens the production of castor bean. Due to the lack of resistance genes, the breeding for root rot resistance in castor bean was seriously restricted. In order to mine resistant resources and establish resistant molecular markers, the phenotypic and molecular marker identification was performed on the disease resistance of 252 castor accessions in this study. The results were as follows: (1) Irrigating roots with the conidia suspension of 1×106 spores ∙ mL-1 was an effective inoculation method. The 5-grade evaluation method based on the days of wilt after inoculation could be used as the criteria to evaluate the resistant level of accessions objectively. (2) According to the criteria, the resistance of 252 accessions were divided into five grades from high to low, among which grade 1 was high resistance and grade 2 was moderate resistance. The number of accessions with different grades from 1 to 5 were 105, 25, 33, 31 and 58 respectively, accounting for 42%, 10%, 13%, 12% and 23% respectively. 130 resistant accessions were identified, of which 105 were high resistance and 25 were moderate resistance. (3) The proportion of resistant accessions in wild accessions (66%) was much higher than that in cultivated accessions (35%). Among wild accessions from South China, 69% were resistant accessions, and 60% were high resistance accessions. It is strongly suggested that the research and utilization of wild accessions, especially the wild accessions in South China, should be an important direction of resistance breeding in the future. (4) 8 SSR markers associated with the resistance were preliminarily established. Although different resistant accessions carried different marker or marker combination, most of them carried 3 to 4 of the above markers, therefore, they can be used as resistant molecular markers for assisted selection. The results of this study provide an effective method and evaluation criteria for root rot resistance identification, screen out a number of resistance genetic resources urgently needed in breeding, and preliminarily establish the SSR markers available for assisted selection, which lay an important foundation for resistance breeding of castor bean root rot.
Hits
Hits
Downloads
Comment