检索结果-内蒙古大学图书馆

7th International Conference on Machine Learning and Natural language Processing, MLNLP 2024

作者： Luo, Yuhao Zhang, Zhengkun Zhang, Yao Xie, Zheng Wang, Jun Jatowt, Adam Yang, Zhenglu Nankai University TKLNDST CS China School of Statistics and Data Science Nankai University LPMC KLMDASR LEBPS China College of Mathematics and Statistics Science Ludong University Shandong Key Laboratory of Language Resource Development and Application China University of Innsbruck Austria

ISBN: (纸本)9798350354973

Event Argument Extraction (EAE) facilitates comprehension of the text related to an event by extracting and analyzing event arguments. The superiority of previous studies typically rises from the abundance of high-quality event annotations, which is labor-intensive to produce and hard to satisfy by reality. In this paper, we approach data-scarce EAE in both low-resource and few-shot scenarios, which have far-reaching implications for practice. Specifically, we propose a model called Dynamic Modular Prompt Tuning based on Slot-Transfer (DAMPT), which dispenses with any manual effort usually required in existing methods. DAMPT turns to large-scale language models to generate dynamic modular prompts, which are more adaptable than the static ones manually given by experts. Furthermore, DAMPT incorporates a prompt-tuning algorithm called slot-transfer to facilitate event-specific knowledge transfer. An extensive experimental evaluation validates the effectiveness and generalization ability of DAMPT in data-scarce scenarios. © 2024 IEEE.

关键词： Modeling languages

来源：评论

学校读者我要写书评

暂无评论

Can We Learn Question, Answer, and Distractors All from An Image? A New Task for Multiple-Choice Visual Question Answering 30

Can We Learn Question, Answer, and Distractors All from An I...

引用

Joint 30th International Conference on Computational Linguistics and 14th International Conference on language resources and Evaluation, LREC-COLING 2024

作者： Ding, Wenjian Zhang, Yao Wang, Jun Jatowt, Adam Yang, Zhenglu TKLNDST CS Nankai University China School of Statistics and Data Science LPMC KLMDASR LEBPS Nankai University China Shandong Key Laboratory of Language Resource Development and Application College of Mathematics and Statistics Science Ludong University China University of Innsbruck Austria

ISBN: (纸本)9782493814104

Multiple-choice visual question answering (MC VQA) requires an answer picked from a list of distractors, based on a question and an image. This research has attracted wide interests from the fields of visual question answering, visual question generation, and visual distractor generation. However, these fields still stay in their own territories, and how to jointly generate meaningful questions, correct answers, and challenging distractors remains unexplored. In this paper, we introduce a novel task, Visual Question-Answer-Distractors Generation (VQADG), which can bridge this research gap as well as take as a cornerstone to promote existing VQA models. Specific to the VQADG task, we present a novel framework consisting of a vision-and-language model to encode the given image and generate QADs jointly, and contrastive learning to ensure the consistency of the generated question, answer, and distractors. Empirical evaluations on the benchmark dataset validate the performance of our model in the VQADG task. © 2024 ELRA language resource Association: CC BY-NC 4.0.

关键词： Benchmarking

来源：评论

学校读者我要写书评

暂无评论

Multimodal language Data and Platform Construction for LCTLs Teaching 23

Multimodal Language Data and Platform Construction for LCTLs...

引用

15th International Conference on Machine Learning and Computing, ICMLC 2023

作者： Wang, Yuan Jin, Kai Liu, Wuying Shandong Key Laboratory of Language Resources Development and Application Ludong University China School of Foreign Languages Qilu University of Technology China Laboratory of Language Engineering and Computing Guangdong University of Foreign Studies China

ISBN: (纸本)9781450398411

The progress of artificial intelligence technology has brought great opportunities for the development of Less Commonly Taught languages (LCTLs) teaching. However, the scarcity of LCTLs resources still plagues the current teaching platform. This paper is committed to building multimodal language data including foreign language vocabulary and translation, bilingual example sentences, phonetic symbols and audio, morphological similarity recommendation, semantic similarity recommendation, etc., and using full-text index and Web Service technology to build a new LCTLs teaching platform. The detailed steps are: (1) acquiring and processing multimodal language data through web pages such as TED and Opensubtitles;(2) constructing the inverted indexes used the HashMap and storing the index files in the cache of the platform, which avoids the heavy database intervention;(3) using the classical Levenshtein distance in the morphological similar words and semantic similar words recommendation algorithm. The platform implemented based on the above data and technology finally supports Japanese, Italian, Indonesian, Malaysian, Vietnamese, Filipino, and Myanmar seven languages. © 2023 ACM.

关键词： Websites

来源：评论

学校读者我要写书评

暂无评论

RETHINKING CNN'S GENERALIZATION TO BACKDOOR ATTACK FROM FREQUENCY DOMAIN 12

RETHINKING CNN'S GENERALIZATION TO BACKDOOR ATTACK FROM FREQ...

引用

12th International Conference on Learning Representations, ICLR 2024

作者： Rao, Quanrui Wang, Lin Liu, Wuying Shandong Key Laboratory of Language Resources Development and Application Ludong University China School of Information and Electrical Engineering Ludong University China Xianda College of Economics and Humanities Shanghai International Studies University China

Convolutional neural network (CNN) is easily affected by backdoor injections, whose models perform normally on clean samples but produce specific outputs on poisoned ones. Most of the existing studies have focused on the effect of trigger feature changes of poisoned samples on model generalization in spatial domain. We focus on the mechanism of CNN memorize poisoned samples in frequency domain, and find that CNN generate generalization to poisoned samples by memorizing the frequency domain distribution of trigger changes. We also investigated the assistance provided by perturbations generated at different frequencies to the generalization, and explore the influence of trigger perturbations in different frequency domain components on the generalization of poisoned models from visible and invisible backdoor attacks. We prove that high-frequency components are more susceptible to perturbations than low-frequency components. Based on the above fundings, we propose a universal invisible strategy for visible triggers, which can achieve trigger invisibility while maintaining raw attack performance. We also design a novel frequency domain backdoor attack method based on low-frequency semantic information, which can achieve 100% attack accuracy on multiple models and multiple datasets, and can bypass multiple defenses. © 2024 12th International Conference on Learning Representations, ICLR 2024. All rights reserved.

关键词： Semantics

来源：评论

学校读者我要写书评

暂无评论

Comparative Analysis of Para-social Interaction among Multiple Luxury Brands 4

Comparative Analysis of Para-social Interaction among Multip...

引用

4th International Conference on Natural language Processing, ICNLP 2022

作者： Yu, Dandan Liu, Wuying Ludong University Shandong Key Laboratory of Language Resources Development and Application Yantai264025 China Guangdong University of Foreign Studies Laboratory of Language Engineering and Computing Guangzhou510420 China

ISBN: (数字)9781665495448

ISBN: (纸本)9781665495448

The luxury industry is good at using the most advanced science and technology to promote product sales. With the recent rise of virtual spokespersons, more and more virtual spokespersons are widely used in luxury marketing. We studied the new advertising mode based on virtual spokespersons, and analyzed how luxury brands use virtual spokespersons to establish para-social interactions with their audiences. We also analyzed the data of likes, reposts and comments on the official Weibo of luxury brands such as Valentino, Dior, Ferragamo, Louis Vuitton, and March Jacobs that have published advertising Weibos which contain virtual spokespersons and do not include celebrity spokespersons and various marketing campaign. Finally, according to the conclusion of our analysis, we propose development suggestions for the promotion strategy of luxury brands' virtual spokespersons in Weibo or other media, and the way to establish para-social interaction between luxury brands and their audiences. © 2022 IEEE.

关键词： Marketing

来源：评论

学校读者我要写书评

暂无评论

Morphological and Semantic Evaluation of Ancient Chinese Machine Translation 1

Morphological and Semantic Evaluation of Ancient Chinese Mac...

引用

1st Workshop on Ancient language Processing, ALP 2023

作者： Jin, Kai Zhao, Dan Liu, Wuying School of Foreign Languages Qilu University of Technology Shandong Jinan250353 China Shandong Key Laboratory of Language Resources Development and Application Ludong University Shandong Yantai264025 China Center for Linguistics and Applied Linguistics Guangdong University of Foreign Studies Guangdong Guangzhou510420 China

ISBN: (纸本)9789544520878

Machine translation (MT) of ancient Chinese texts presents unique challenges due to the complex grammatical structures, cultural nuances, and polysemy of the language. This paper focuses on evaluating the translation quality of different platforms for ancient Chinese texts using The Analects as a case study. The evaluation is conducted using the BLEU, LMS, and ESS metrics, and the platforms compared include three machine translation platforms (Baidu Translate, Bing Microsoft Translator, and DeepL), and one language generation model ChatGPT that can engage in translation endeavors. Results show that Baidu performs the best, surpassing the other platforms in all three metrics, while ChatGPT ranks second and demonstrates unique advantages. The translations generated by ChatGPT are deemed highly valuable as references. The study contributes to understanding the challenges of MT for ancient Chinese texts and provides insights for users and researchers in this field. It also highlights the importance of considering specific domain requirements when evaluating MT systems. © RANLP-ALP 2023 - Proceedings of the Ancient language Processing Workshop, associated with 14th International Conference on Recent Advances in Natural language Processing.

关键词： Machine translation

来源：评论

学校读者我要写书评

暂无评论

Assessing spatiotemporal risks of nonpoint source pollution via soil erosion: a coastal case in the Yellow River Delta, China

引用

Environmental Science and Pollution Research 2024年第23期31卷 34569-34587页

作者： Wang, Youxiao Huang, Chong Liu, Gaohuan Zhao, Zhonghe Li, He Sun, Yingjun School of Surveying and Geo-Informatics Shandong Jianzhu University Jinan250101 China State Key Laboratory of Resources and Environmental Information System Institute of Geographic Sciences and Natural Resources Research Chinese Academy of Sciences Beijing100101 China Jiangsu Center for Collaborative Innovation in Geographical Information Resource Development and Application Nanjing210023 China Agricultural Information Institute Chinese Academy of Agricultural Sciences Beijing100081 China

Nonpoint source pollution (NPSP) has always been the dominant threat to regional waters. Based on empirical models of the revised universal soil loss equation and the phosphorus index, an NPSP risk assessment model denoted as SL-NPSRI was developed. The surface soil pollutant loss was estimated by simulating the rain-runoff topographic process, and the influence of path attenuation was quantified. A case study in the Yellow River Delta and corresponding field surveys of soil pollutants and water quality showed that the established model can be applied to evaluate the spatial heterogeneity of NPSP. NPSP usually occurs during high-intensity rainfall periods and in larger estuaries. Summer rainfall increased pollutant transport into the sea from late July to mid-August and caused estuarine dilution. Higher NPSP risks often correspond to coastal areas with lower vegetation coverage, higher soil erodibility, and higher soil pollutant concentrations. Agricultural NPSP originating from cropland significantly increase the pollutant fluxes. Therefore, area-specific land use management and vegetation coverage improvement, and temporal-specific strategies can be explored for NPSP control during source-transport hydrological processes. This research provides a novel insight for coastal NPSP simulations by comprehensively analyzing the soil erosion process and its associated pollutant loss effects, which can be useful for targeted spatiotemporal solutions. Graphical Abstract: (Figure presented.) © The Author(s), under exclusive licence to Springer-Verlag GmbH Germany, part of Springer Nature 2024.

关键词： Erosion

来源：评论

学校读者我要写书评

暂无评论

Morphological Semantic Ensemble Filtering of Massive Sentence Pairs for Neural Machine Translation 12th

Morphological Semantic Ensemble Filtering of Massive Sentenc...

引用

12th CCF Conference on BigData, BigData 2024

作者： Wang, Lin Liu, Wuying Xianda College of Economics and Humanities Shanghai International Studies University Shanghai200083 China Shandong Key Laboratory of Language Resources Development and Application Ludong University Yantai264025 China

ISBN: (纸本)9789819610235

The dazzling brilliance of deep learning makes it a consensus that "give me large-scale high-quality sentence pairs to train on, I can obtain a high-performance neural machine translation model." The current ever-increasing productivity of human language data greatly reduces the scarcity of sentence pairs, but the explosion of language data also brings potential problems such as uneven translation quality of sentences. We address the issue of sentence pair filtering based on translation quality, explore a new method of morphological semantic ensemble filtering, which can fully leverage the efficiency of morphological measurement derived from Levenshtein editing distance and the accuracy of semantic measurement derived from pretrained models, and achieve an efficient estimation of sentence pair translation quality taking into account both morphology and semantics. We first conduct morphological filtering, semantic filtering, and morphological semantic ensemble filtering experiments on the datasets of 17 language pairs respectively, and then use the filtered sentence pairs to enhance the retraining of the 17 neural machine translation models. Experimental results show that all three filtered results can significantly improve the performance of neural machine translation models, among which, the morphological semantic ensemble filtering has the best effect, improving 2–3 BLEU points compared with the other two methods. Experimental results clarify that our new method can use morphological computing to strengthen perfect homonym data, while can use semantic computing to strengthen shaped synonym data. The corresponding implementation is an industrial-level straightforward and efficient algorithm. © The Author(s), under exclusive license to Springer Nature Singapore Pte Ltd. 2025.

关键词： Wiener filtering

来源：评论

学校读者我要写书评

暂无评论

Data-Scarce Event Argument Extraction: A Dynamic Modular Prompt Tuning Model Based on Slot Transfer

Data-Scarce Event Argument Extraction: A Dynamic Modular Pro...

引用

Machine Learning and Natural language Processing (MLNLP), International Conference on

作者： Yuhao Luo Zhengkun Zhang Yao Zhang Zheng Xie Jun Wang Adam Jatowt Zhenglu Yang TKLNDST CS Nankai University LPMC KLMDASR & LEBPS School of Statistics and Data Science Nankai University Shandong Key Laboratory of Language Resource Development and Application College of Mathematics and Statistics Science Ludong University University of Innsbruck Austria

ISBN: (数字)9798350354973

ISBN: (纸本)9798350354980

关键词： Adaptation models Annotations Heuristic algorithms Manuals Machine learning Transformers Natural language processing Tuning Knowledge transfer

来源：评论

学校读者我要写书评

暂无评论

Quantitative Analysis of Land Degradation in Mongolia from the Perspective of Geographical Zone

引用

Journal of resources and Ecology 2024年第5期15卷 1185-1192页

作者： HAN Tengfei WANG Juanle HAN Baomin School of Civil Engineering and Geomatics Shandong University of TechnologyZiboShandong 255049China State Key Laboratory of Resources and Environmental Information Systems Institute of Geographic Sciences and Natural Resources ResearchChineseAcademy of SciencesBeijing 100101China Jiangsu Center for Collaborative Innovation in Geographical Information Resource Development and Application Nanjing 210023China

Natural and anthropogenic disturbances accelerate land degradation(LD)in arid,semi-arid,and dry sub-humid areas,leading to reduced land quality and productivity,loss of biodiversity,degradation of ecosystem services,and a decline in the quality of life of local *** address this issue,the United Nations Convention to Combat Desertification(UNCCD)has set a target for LD neutrality(LDN).However,quantifying and comparing the status of LD at global or regional scales remains challenging due to the lack of coherent quantitative methods and *** this study,we focused on Mongolia,a region with significant LD problems,to examine patterns of LD and changes from 2015 to 2020,accounting for regional *** was used,as recommended by the *** main findings are as follows:(1)Overall,the degraded land area in Mongolia accounted for 12.11%of the total land area,predominantly located in the southwest desert and desert steppe,gradually spreading to the northeast steppe.(2)The areas showing improvement in the land productivity index and degradation were 17.62%and 11.79%,respectively,with the most severely degraded areas concentrated in the southern desert and desert steppe regions.(3)The areas of improvement and degradation in the land cover index were 1.80%and 0.16%,respectively,with degraded areas scattered across regions of steppe,high mountains,and mountain taiga.(4)The areas of improvement and degradation in the land organic carbon index were 1.54%and 0.22%,respectively,with degradation primarily observed in adjacent areas of mountain taiga,steppe,and desert steppe.(5)The improved area(2.999×10^(5)km^(2))of LDN are more than the degraded area(1.895×10^(5)km^(2)),indicating a positive trend toward LDN in Mongolia.

关键词： geographical zone land degradation land degradation neutrality Mongolia Trends.Earth

来源：评论

学校读者我要写书评

暂无评论

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案：

请选择收藏分类：

通借通还

建议与咨询 留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

时间限定

文献类型

馆藏选择

核心期刊

语言

文献类型

帮助

文字说明：

检索规则说明：

检索范例：

分类表

所选分类

限定检索结果

文献类型

馆藏范围

日期分布

学科分类号

主题

机构

作者

语言

请选择保存的检索档案： 新增检索档案 确定 取消

请选择收藏分类： 新增自定义分类 确定 取消

通借通还

建议与咨询留下您的常用邮箱和电话号码，以便我们向您反馈解决方案和替代方法

请选择保存的检索档案：

请选择收藏分类：