Hasil untuk "Chinese language and literature"

Richárd Gábor Gottner

This paper aims to examine the self-defence techniques of the Kyokushinkai karate style, with a primary focus on one of its popular legends. According to the myth, the founder Ōyama Masutatsu had extensive training in the art of Daitō-ryū aikijūjutsu (the ancestor art of modern aikidō) under a master called Yoshida Kōtarō and received the ‘license of full transmission’ (menkyo kaiden) scroll of the Daitō-ryū school. The basis of this research is a high-resolution photocopy of the document (unfortunately due of copyright concerns, the picture of the scroll could not be featured in this paper; however a complete transcription and translation are included in the appendices). The paper dwells into the historical connections between the arts of Kyokuhsinkai, Daitō-ryū aikijūjutsu, and aikidō supported by circumstantial evidence and gives a brief technical comparation of the self-defence techniques of Kyokushinkai and the curricula of Daitō-ryū and aikidō. This is followed by a detailed analysis of the scroll itself, examining its structure and terminology and some techniques included in the scroll that are possibly featured in the Kyokushinkai curriculum, with special attention to stick and parasol techniques. The research gives a full analysis of Kyokushinkai self-defence techniques, which can be found in the original and revised editions of What is Karate?, This is Karate, Advanced Karate, and Mas Oyama’s Essential Karate, written by the founder of Ōyama Masutatsu himself, as well as Kyokushin Karate Self-Defense Techniques by high-ranking Kyokuhsinkai practitioner Bobby Lowe. The self-defence techniques found in the aforementioned texts show many similarities to the techniques published in various jūdō, aikidō, and aikijūjutsu books. Considering the historical facts (based on circumstantial evidence) that Ōyama Masutatsu trained in jūdō in the gym of Sone Kōzō, was a direct student of Daitō-ryū instructor Yoshida Kōtarō, and personally knew (allegedly even trained under) aikidō master Shioda Gōzō (founder of the Yōshinkan style of aikidō), it is highly likely that the parallels between the various self-defence techniques of Kyokushinkai karate, Kōdōkan jūdō, Yōshinkan aikidō, and Daitō-ryū aikijūjutsu are not merely coincidence. However, a thorough examination of the supposed Daitō-ryū scroll that Ōyama received from Yoshida Kōtarō revealed that the document itself is not a menkyo kaiden (license of full transmission) certificate. Furthermore, compared to the extant official scrolls of the Daitō-ryū tradition (Hiden mokuroku, Hiden okugi no koto), while the structure is similar, the scroll of Ōyama is certainly not a conventional transmission document of the Daitō-ryū school. The scroll that Yoshida granted to Ōyama is titled Yoshida-shiki sutekki-parasoru goshinjutsu mokuroku (Yoshida-style stick-parasol self-defence scroll), which (at the time of this research) appears to be a ‘one-of-akind’ transmission document on its own, officially signed by ‘The founder of Yoshida-style stickparasol self-defence, Daitō-ryū jūjutsu representative instructor, Yoshida Kōtarō’. The techniques listed in the scroll are unique in their terminology, and since there is no further information about the execution of these techniques, there is only a limited, assumption-based possibility to compare these techniques to those of the Daitō-ryū aikijūjutsu curriculum. For the same reason, it is also nearly impossible to certainly conclude whether the techniques from the scroll are featured as self-defence techniques in Ōyama’s karate books. Certain techniques found in the books could be matched with some technical names from the scroll; however, these connections cannot be stated with complete certainty. The research concluded in this paper answers the questions of whether Ōyama Masutatsu received a ‘license of full transmission’ in the Daitō-ryū school and whether there is in fact a possible connection between the Kyokushinkai self-defence techniques and the Daitō-ryū aikijūjutsu curriculum. While there are certainly Daitō-ryū techniques (among others) similar to those of the Kyokushinkai self-defence curriculum, and Ōyama without a doubt was a student of a Daitō-ryū representative instructor, he never received an official transmission license in the art of Daitō-ryū aikijūjutsu.

Zhengfang Chen, Jing Zhao, Xiaoxuan Fan et al.

IntroductionVariants in the SR-related C-terminal domain-Associated factor 4 (SCAF4) gene are linked to Fliedner-Zweier syndrome (FZS), which presents with diverse symptoms, including mild intellectual disability, seizures, behavioral abnormalities, and various skeletal and structural anomalies. However, there is a paucity of cases describing genotypes and clinical features.Case presentationWe present the case of a 4-year and seven-month-old Chinese boy displaying intellectual impairment, language development disorder, behavioral abnormalities, and distinct facial features. Whole exome sequencing (WES) identified a heterozygous nonsense mutation, c.1693C>T (p.Arg565*), located in exon 14 of the SCAF4 gene (NM_020706). Sanger sequencing confirmed paternal inheritance of this mutation. RNA sequencing from the patient demonstrated widespread transcriptional dysregulation, reinforcing the role of SCAF4 dysfunction in impaired transcription and neurodevelopmental disorders. This mutation is novel, not previously recorded in databases such as GnomAD or dbSNP, nor reported in existing literature.ConclusionWe reviewed the clinical features of the patients reported in the literature with mutations in SCAF4 gene and described the case of a Chinese patient with this mutation. This case underscores the critical need for continued exploration of genotype-phenotype correlations, enhancing our understanding of the diverse manifestations of Fliedner-Zweier syndrome and informing future diagnostic and therapeutic strategies.

Lynda Ben Hazem

To create standards for foreign language instruction in grades K–12, a coalition of four national language organizations—the Council on Foreign Language Teaching in the United States, the Association of French Teachers in the United States, the Association of German Teachers in the United States, and the Association of Spanish and Portuguese Teachers in the United States—was granted funding in 1993. To define content standards for foreign language education, an eleven-member task force representing various languages, program models, levels of instruction, and geographic areas was assigned. The task force disseminated its findings to the general public and the larger profession at every stage of its formation. The resulting document, which defines and defines the role of education, represents an unprecedented agreement among educators, business leaders, the government, and the community. As a multicultural country, the People's Republic of Tunisia began Chinese language education in the 1970s and offered a three-year undergraduate program in Chinese language at Carthage University’s Higher Language Institute of Tunis in Tunisia. The teachers in this major include locals, non-locals, and Chinese. All the teachers have designed a teaching plan together that can provide students with various aspects of Chinese language education. Since Chinese language teaching for said students is a base education process (meaning most of the students do not have a prior Chinese language education), this article will use the 5C plan for Chinese language learning goals in primary and secondary schools across the United States to separately define the content and objectives of this teaching plan, therefore highlighting it’s strengths and potential weaknesses to help better the Chinese education in Tunisia, aiding in the deepening of Sino-Tunisian relations.

Szandra Ésik

Szandra Ésik's review of the Chinese–Hungarian Dictionary (2017) edited by Huba Bartos and Imre Hamar, and the Hungarian–Chinese Dictionary (2024) edited by Huba Bartos, Imre Hamar and Melinda Pap.

Shanshan Wang, Junchao Wu, Fengying Ye et al.

The rapid development of advanced large language models (LLMs) has made AI-generated text indistinguishable from human-written text. Previous work on detecting AI-generated text has made effective progress, but has not involved modern Chinese poetry. Due to the distinctive characteristics of modern Chinese poetry, it is difficult to identify whether a poem originated from humans or AI. The proliferation of AI-generated modern Chinese poetry has significantly disrupted the poetry ecosystem. Based on the urgency of identifying AI-generated poetry in the real Chinese world, this paper proposes a novel benchmark for detecting LLMs-generated modern Chinese poetry. We first construct a high-quality dataset, which includes both 800 poems written by six professional poets and 41,600 poems generated by four mainstream LLMs. Subsequently, we conduct systematic performance assessments of six detectors on this dataset. Experimental results demonstrate that current detectors cannot be used as reliable tools to detect modern Chinese poems generated by LLMs. The most difficult poetic features to detect are intrinsic qualities, especially style. The detection results verify the effectiveness and necessity of our proposed benchmark. Our work lays a foundation for future detection of AI-generated poetry.

Jin Jiani

In recent years, the Chinese government has vigorously promoted the rural revitalization strategy, aiming to achieve a balanced development of urban and rural social economy and promote common prosperity. The key to rural revitalization lies in talents. With the continuous improvement of Zhoushan's internationalization and opening up, Zhoushan's local foreign language talent pool is facing the dual pressure of quantity and quality, including the inadaptation of the training mode of business English education talents, the serious loss of high-quality talents in business English education and the uneven distribution of educational resources. Based on the current situation of business English education in Zhoushan City, this paper adopts literature research, expert interview, questionnaire survey, field investigation and other research methods. According to the results of questionnaire analysis and regional characteristics, this paper puts forward strategies to improve the business English education system in colleges and universities, strengthen the cooperation between colleges and enterprises, and promote the deep integration of information technology and business English teaching, hoping to provide new development opportunities for business English education to help rural revitalization.

Li Xiangning

In order to more intuitively analyze the research status and research trends of motivational factors in second language acquisition, this paper summarizes and combs through all the relevant articles included in the CNKI database from 2009-2024 with the help of the Citespace tool, and analyzes the number of articles, research hotspots, and research trends in the field through keyword co-occurrence, cluster analysis, keyword time-zone diagrams, and other visualization methods. It is found that (1) the number of articles published on motivational factors in second language acquisition is generally cyclical, with fluctuations, but overall the number of articles is relatively stable; (2) the research hotspots in this field are centered on the research hotspots of English language learning, dynamics, influencing factors, and online learning; and (3) since 2020, the study of international postures, changes in motivational dynamics and retrospectives is emerging as a new direction in the study of motivational factors.

Xudong Lu, Qi Liu, Yuhui Xu et al.

A pivotal advancement in the progress of large language models (LLMs) is the emergence of the Mixture-of-Experts (MoE) LLMs. Compared to traditional LLMs, MoE LLMs can achieve higher performance with fewer parameters, but it is still hard to deploy them due to their immense parameter sizes. Different from previous weight pruning methods that rely on specifically designed hardware, this paper mainly aims to enhance the deployment efficiency of MoE LLMs by introducing plug-and-play expert-level sparsification techniques. Specifically, we propose, for the first time to our best knowledge, post-training approaches for task-agnostic and task-specific expert pruning and skipping of MoE LLMs, tailored to improve deployment efficiency while maintaining model performance across a wide range of tasks. Extensive experiments show that our proposed methods can simultaneously reduce model sizes and increase the inference speed, while maintaining satisfactory performance. Data and code will be available at https://github.com/Lucky-Lance/Expert_Sparsity.

Xuechun Wang, Rodney Beard, Rohitash Chandra

Machine translation using large language models (LLMs) is having a significant global impact, making communication easier. Mandarin Chinese is the official language used for communication by the government and media in China. In this study, we provide an automated assessment of translation quality of Google Translate with human experts using sentiment and semantic analysis. In order to demonstrate our framework, we select the classic early twentieth-century novel 'The True Story of Ah Q' with selected Mandarin Chinese to English translations. We use Google Translate to translate the given text into English and then conduct a chapter-wise sentiment analysis and semantic analysis to compare the extracted sentiments across the different translations. Our results indicate that the precision of Google Translate differs both in terms of semantic and sentiment analysis when compared to human expert translations. We find that Google Translate is unable to translate some of the specific words or phrases in Chinese, such as Chinese traditional allusions. The mistranslations may be due to lack of contextual significance and historical knowledge of China.

Gihun Lee, Minchan Jeong, Yujin Kim et al.

While learning to align Large Language Models (LLMs) with human preferences has shown remarkable success, aligning these models to meet the diverse user preferences presents further challenges in preserving previous knowledge. This paper examines the impact of personalized preference optimization on LLMs, revealing that the extent of knowledge loss varies significantly with preference heterogeneity. Although previous approaches have utilized the KL constraint between the reference model and the policy model, we observe that they fail to maintain general knowledge and alignment when facing personalized preferences. To this end, we introduce Base-Anchored Preference Optimization (BAPO), a simple yet effective approach that utilizes the initial responses of reference model to mitigate forgetting while accommodating personalized alignment. BAPO effectively adapts to diverse user preferences while minimally affecting global knowledge or general alignment. Our experiments demonstrate the efficacy of BAPO in various setups.

Yuanze Ji, Bobo Li, Jun Zhou et al.

Multimodal Named Entity Recognition (MNER) is a pivotal task designed to extract named entities from text with the support of pertinent images. Nonetheless, a notable paucity of data for Chinese MNER has considerably impeded the progress of this natural language processing task within the Chinese domain. Consequently, in this study, we compile a Chinese Multimodal NER dataset (CMNER) utilizing data sourced from Weibo, China's largest social media platform. Our dataset encompasses 5,000 Weibo posts paired with 18,326 corresponding images. The entities are classified into four distinct categories: person, location, organization, and miscellaneous. We perform baseline experiments on CMNER, and the outcomes underscore the effectiveness of incorporating images for NER. Furthermore, we conduct cross-lingual experiments on the publicly available English MNER dataset (Twitter2015), and the results substantiate our hypothesis that Chinese and English multimodal NER data can mutually enhance the performance of the NER model.

GUO Pengrui, WEN Tingxiao

[Purpose/Significance] This article is aimed to explore the impact of artificial intelligence generation technologies such as large language models (LLMs) on users' information retrieval behavior and to suggest ideas for information retrieval systems and information resource construction. In this way, it provides insights into and references for the future establishment of the artificial intelligence generated content (AIGC) information platform with Chinese characteristics as well as the information literacy education system. [Method/Process] In the field of library intelligence, with the wide application of AI technology in information service work, LLMs represented by ChatGPT have also become a hot topic of discussion. Taking the booming development of LLMs such as ChatGPT as background, we analyzed the impact of the increasing popularity of this technology on information retrieval systems and user retrieval behavior from the perspective of user information behavior by combining the technical features of LLMs with the characteristics of existing products. Literature survey and empirical analysis were used. [Results/Conclusions] The use of LLMs as information retrieval systems has unparalleled advantages over traditional products. These advantages include the ability to understand and process natural language queries, generate relevant and context-specific responses, and interact with users in a more human-like way. The application of LLMs in information retrieval systems has the potential to transform the way users search for information, influence the underlying logic, action priorities, and retrieval expectations of user information retrieval behavior. However, the existing shortcomings of LLMs in terms of reliability and accuracy still make it difficult for them to replace traditional information retrieval methods immediately. Language models may not always provide accurate and reliable answers, especially when dealing with complex or domain-specific queries. Additionally, LLMs may struggle to understand and process contextual information effectively, leading to limitations in their ability to extract relevant and context-aware insights. It is recommended to pay attention to this technology in the construction of information retrieval systems and information resources, and to explore the combination of LLMs and information services in order to cope with the changes in future user information needs and to further make full use of the value of existing information resources. Limited by the lack of expertise in the field of AI and the fact that LLMs are not yet widely used in practice in China, the research findings are only a reflection and exploration of the impact of LLMs on users' information behavior.

Hanwool Lee, Jonghyun Choi, Sohyeon Kwon et al.

This paper presents our participation in the FinNLP-2023 shared task on multi-lingual environmental, social, and corporate governance issue identification (ML-ESG). The task's objective is to classify news articles based on the 35 ESG key issues defined by the MSCI ESG rating guidelines. Our approach focuses on the English and French subtasks, employing the CerebrasGPT, OPT, and Pythia models, along with the zero-shot and GPT3Mix Augmentation techniques. We utilize various encoder models, such as RoBERTa, DeBERTa, and FinBERT, subjecting them to knowledge distillation and additional training. Our approach yielded exceptional results, securing the first position in the English text subtask with F1-score 0.69 and the second position in the French text subtask with F1-score 0.78. These outcomes underscore the effectiveness of our methodology in identifying ESG issues in news articles across different languages. Our findings contribute to the exploration of ESG topics and highlight the potential of leveraging advanced language models for ESG issue identification.

Hannah Rose Kirk, Andrew M. Bean, Bertie Vidgen et al.

Human feedback is increasingly used to steer the behaviours of Large Language Models (LLMs). However, it is unclear how to collect and incorporate feedback in a way that is efficient, effective and unbiased, especially for highly subjective human preferences and values. In this paper, we survey existing approaches for learning from human feedback, drawing on 95 papers primarily from the ACL and arXiv repositories.First, we summarise the past, pre-LLM trends for integrating human feedback into language models. Second, we give an overview of present techniques and practices, as well as the motivations for using feedback; conceptual frameworks for defining values and preferences; and how feedback is collected and from whom. Finally, we encourage a better future of feedback learning in LLMs by raising five unresolved conceptual and practical challenges.

P. Saravanan, S. Rajeswari, J. A. Kumar et al.

H. Vonková, Jane Jones, Angie Moore et al.

Xinyi Liu, Ying Zeng, Juan He et al.

Purpose The purpose of this study is to present a holistic review of journal articles on value cocreation in the tourism and hospitality field. By distinguishing similarities and differences in the contexts of China and other regions, the authors identify targeted research directions. Design/methodology/approach This study collected 438 English articles and 31 Chinese articles on value cocreation in tourism and hospitality published in the Web of Science, Scopus and China National Knowledge Infrastructure online databases. After analyzing the study location of all articles in the sample, the authors identified 79 English-language articles that adopted China as the context of the study. The first data set included non-China-focused publications and the second data set included China-focused publications. A comparative bibliometric methodology was conducted. Findings The results indicate similarities and differences between the two data sets in terms of research methods, topics and future research directions. Seven major concepts were identified in the first data set: hospitality value cocreation; the value generation process; key stakeholders; outcome variables; the application of service-dominant logic; peer-to-peer accommodation and destination management. Four research focuses were identified in the second data set: the application of customer-dominant logic; the value generation process; resource integration and the cocreation experience. Research limitations/implications This research contributes to the understanding of tourism and hospitality cocreation systems through literature analysis and encourages future research on the nexus of the Chinese and global contexts. This systematic investigation extends the concept of value cocreation by integrating its implementation process and value categories and further evaluates the consequences and benefits of value cocreation, which helps clarify how winners and losers can adjust their strategies accordingly. Originality/value The originality of this paper lies in the adopted comparative bibliometric analysis which comprehensively reviewed and compared the characteristics of value cocreation research across two contexts. In addition, the study dialectically assesses Chinese tourism cocreation phenomena and their impacts from a global perspective.

Hui Yang

Yongrong Wu, Xujun You, Xujun You et al.

Background: Drug resistance is the major cause of increasing mortality in prostate cancer (PCa). Therefore, it an urgent to develop more effective therapeutic agents for PCa treatment. Xihuang pills (XHP) have been recorded as the efficient anti-tumor formula in ancient Chinese medical literature, which has been utilized in several types of cancers nowadays. However, the effect protective role of XHP on the PCa and its underlying mechanisms are still unclear.Methods: The active ingredients of XHP were obtained from the Traditional Chinese Medicine Systems Pharmacology Database and Analysis Platform (TCMSP) and BATMAN-TCM. The potential targets of PCa were acquired from the Gene Cards and OMIM databases. R language and Perl language program were utilized to clarify the interaction between the PCa-related targets and the potential targets of XHP. The potential targets of XHP for prostate cancer were gathered from the Gene ontology and KEGG pathway. Furthermore, cell proliferation assays were verified by PC3 and LNCaP cells. The efficacy and potential mechanism tests were confirmed by the PCa PC3 cells and mice subcutaneous transplantation. The effects of PI3K/Akt/mTOR-related proteins on proliferation, apoptosis, and cell cycle of PCa cells were measured by the Cell Counting Kit-8(CCK8), TUNEL assay, real-time quantitative reverse transcription PCR (QRT-PCR), and Western Blotting, respectively.Results: The active components of four traditional Chinese medicines in XHP were searched on the TCMSP and Batman TCM database. The biological active components of XHP were obtained as OB ≥30% and DL ≥0.18. The analysis of gene ontology and KEGG pathway identified the PI3K/Akt/mTOR signaling pathway as the XHP-associated pathway. Collectively, the results of in vitro and in vivo experiments showed that XHP had the effect of inhibiting on the proliferation of PC3 and LNCaP cells. XHP promoted the apoptosis and restrained the cell cycle and invasion of the PC3 cells and subcutaneous transplantation. Meanwhile, the suppression of XHP on the level of expression of PI3K, Akt, and mTOR-pathway-related pathway proteins has been identified in a dose-dependent manner.Conclusion: PI3K/Akt/mTOR pathway-related pathway proteins were confirmed as the potential XHP-associated targets for PCa. XHP can suppress the proliferation of prostate cancer via inhibitions of the PI3K/Akt/mTOR pathway.

Ying Yin, Ivan Habernal

Pre-training large transformer models with in-domain data improves domain adaptation and helps gain performance on the domain-specific downstream tasks. However, sharing models pre-trained on potentially sensitive data is prone to adversarial privacy attacks. In this paper, we asked to which extent we can guarantee privacy of pre-training data and, at the same time, achieve better downstream performance on legal tasks without the need of additional labeled data. We extensively experiment with scalable self-supervised learning of transformer models under the formal paradigm of differential privacy and show that under specific training configurations we can improve downstream performance without sacrifying privacy protection for the in-domain data. Our main contribution is utilizing differential privacy for large-scale pre-training of transformer language models in the legal NLP domain, which, to the best of our knowledge, has not been addressed before.