Hasil untuk "Japanese language and literature"

Filip Kučera, Christoph Mandl, Isao Echizen et al.

Definitions are the foundation for any scientific work, but with a significant increase in publication numbers, gathering definitions relevant to any keyword has become challenging. We therefore introduce SciDef, an LLM-based pipeline for automated definition extraction. We test SciDef on DefExtra & DefSim, novel datasets of human-extracted definitions and definition-pairs' similarity, respectively. Evaluating 16 language models across prompting strategies, we demonstrate that multi-step and DSPy-optimized prompting improve extraction performance. To evaluate extraction, we test various metrics and show that an NLI-based method yields the most reliable results. We show that LLMs are largely able to extract definitions from scientific literature (86.4% of definitions from our test-set); yet future work should focus not just on finding definitions, but on identifying relevant ones, as models tend to over-generate them. Code & datasets are available at https://github.com/Media-Bias-Group/SciDef.

Shigeo Ohkubo

In 1954, following a five-year research period in the U.S., Professor Hideki Yukawa returned to Japan and visited Kochi on March 21 to attend the unveiling ceremony for the first statue of him ever built in Japan, a project initiated by the PTA of Yasu Elementary School in Yasu Town, Kochi Prefecture. By a coincidence of history, just three weeks prior on March 1, the U.S. had conducted a hydrogen bomb test at Bikini Atoll in the Pacific Ocean. Many Japanese fishing boats were operating there at the time and had not been informed in advance. As a result, numerous boats, including the Daigo Fukuryu Maru, were exposed to radiation. Upon his arrival at Kochi Station on the evening of March 21, Yukawa was relentlessly questioned by reporters about the Bikini hydrogen bomb. This was a source of deep anguish for Yukawa, a Japanese physicist who had won the Nobel Prize for his work on "atomic physics." He firmly refused to answer, stating that the topic was "outside the scope of my research." The next evening, at a public lecture in Kochi City on March 22, he again refused to speak about the Bikini hydrogen bomb or nuclear power, stating that he was an amateur in nuclear research and that there were many other experts. However, just four days later, on March 28, after returning to Kyoto, Yukawa drafted his famous essay, "The Turning Point for Humanity and Atomic Power," which was published in a newspaper on March 30. From that point on, he was drawn into the tumultuous issue of the Bikini hydrogen bomb and nuclear power. When did a tormented Yukawa make his decision? This article meticulously reveals, based on historical documents, what led the anguished Yukawa to make such a rapid decision within a single day and what caused the immense change in his mindset overnight.

Masaharu Mizumoto, Dat Nguyen, Zhiheng Han et al.

Benchmark saturation and contamination have obscured genuine advances in reasoning for large language models (LLMs). We introduce NazoNazo Benchmark, a low-cost, renewable test built from Japanese children's riddles that demand insight-based reasoning, or representational shifts rather than knowledge recall. We evaluate 38 frontier LLMs (2023-2025) on 201 riddles and a 120-item human-comparison subset, finding that non-reasoning models average 7.6%, reasoning models 17.6%, and humans ~53% accuracy. Importantly, thought-log analysis reveals that reasoning in Japanese did not necessarily improve accuracy, indicating that language understanding alone is insufficient for insight reasoning. Notably, models sometimes generated correct candidates but failed to endorse them, suggesting weak metacognitive control rather than a lack of knowledge. This "verification failure" indicates that CoT outputs can reflect genuine intermediate reasoning states rather than post-hoc rationalizations. By exposing this metacognitive bottleneck - models' inability to recognize when they are right - the benchmark provides a scalable, cross-linguistic testbed for studying machine insight, confidence calibration, and self-evaluation. NazoNazo Benchmark thus offers not only a fresh challenge to current LLMs but also a concrete target for developing AI metacognitive psychology and enhancing machine Aha! capability.

Cringuta Irina Pelea

The present article examines the digital communication patterns exerted by domestic Japanese companies through Japanese-language blogging practices: main characteristics, current trends, challenges, and future tendencies at the intersection of Japanese corporate culture, communication, and the digital realm. After a concise literature review of Japanese corporate identity and predominant cultural features of Japanese management, we will conduct a qualitative content analysis on a selected corpus of Japanese language corporate blogs to investigate the digital corporate discourse (content, types of public, communication strategies, and techniques) on a national level and to inquire into the extent of the impact exerted by the digital evolution of media platforms and the emergence of new AI tools. The implications of the present research are twofold. First, the database we manually compiled contains only Japanese language sources, thus surpassing the linguistic barrier many Western academics confront when researching Asian countries and media content in Asian languages. Examining the Japanese language digital discourse of domestic companies in a post-pandemic globalized context bridges the unceasing gap between the West and the East, thus decentring the focal point from the Western companies and shifting the attention from English-language digital communication. Second, connecting the Japanese-language business discourse with domestic blogging platforms represents another aspect of particular importance neglected by previous research.

Juliana Buriticá Alzate

This article focuses on depictions of breastfeeding that move beyond idealization and/or realistic depictions of breastfeeding, and do not necessarily elicit an empathic, personal connection with the reader, but rather, I argue that it is via their engagement with Japanese folklore and nature, that they create the necessary distance to critically explore the question of breastfeeding within a feminist framework. I analyze the fantastical treatment of breastfeeding through its fictional representation in two short stories: “Higanbana” (彼岸花 Spider Lilies, 2014; trans. 2014) by Oyamada Hiroko, and “Enoki no isshо̄” (エノキの一生 Enoki, 2019; trans. 2020) by Matsuda Aoko. One of the representative features of the writing of both Oyamada and Matsuda is how they bring together fantastic events within realistic settings, blending realism and science fiction, the human and the non-human, the ordinary and the extraordinary. In the selected stories, both authors play with traditional beliefs and popular stories passed by word of mouth as well as feature personification and imagery of plants, flowers, and trees. This literary quality—problematizing “nature” and folklore while blending fantasy and reality—can be read as a subversive strategy for potential feminist critique. Hence, I read their depiction of breastfeeding in terms of it being fantasy grounded in reality, focusing on unraveling affects of shame and disgust, often associated to breastfeeding, and moving towards a reproductive justice framework.

Yukiko Ishizuki, Tatsuki Kuribayashi, Yuichiroh Matsubayashi et al.

Speakers sometimes omit certain arguments of a predicate in a sentence; such omission is especially frequent in pro-drop languages. This study addresses a question about ellipsis -- what can explain the native speakers' ellipsis decisions? -- motivated by the interest in human discourse processing and writing assistance for this choice. To this end, we first collect large-scale human annotations of whether and why a particular argument should be omitted across over 2,000 data points in the balanced corpus of Japanese, a prototypical pro-drop language. The data indicate that native speakers overall share common criteria for such judgments and further clarify their quantitative characteristics, e.g., the distribution of related linguistic factors in the balanced corpus. Furthermore, the performance of the language model-based argument ellipsis judgment model is examined, and the gap between the systems' prediction and human judgments in specific linguistic aspects is revealed. We hope our fundamental resource encourages further studies on natural human ellipsis judgment.

Masaaki Kimura, Zen Walsh, Takuo Inoue et al.

In the trend towards the globalization of football and the increasing commercialization of professional football clubs, a methodology for calculating the firm value of clubs in non-western countries has yet to be established. This study reviews the valuation methods for the club firm values in Europe and North America and how values are calculated at the time of changing ownership of Japanese clubs and develops regression models with higher explanatory power than before to estimate the more accurate firm value of Japanese football clubs. A review of the existing literature on methods for calculating the firm value of professional sports clubs in Europe and North America, as well as financial statements and registers relating to changes of ownership of Japanese clubs, was conducted. After that, multiple regression analyses were conducted using the firm value of European clubs as the explained variable. From the literature review and the Japanese case studies, it has become clear that European clubs' standard valuation methods are based on revenue and other factors, while in Japan, valuation is based solely on the par value of stocks or net assets. Multiple regression analysis revealed that the firm value of European clubs over the past three years is best explained by revenue or player market value and the number of SNS followers. Two models with high explanatory power were developed. The estimated firm value using the revenue-based formula was higher than the one based on player market value. However, in the J.League, the former was more than three times higher than the latter, while the former was only 1.2 times higher for European clubs. The discrepancy relates to differences in European and J.League clubs' revenues and asset structures. In either formula, the firm value of J.League clubs exceeded the actual transaction price when the change of ownership occurred in the past.

Fangru Lin, Daniel Altshuler, Janet B. Pierrehumbert

Scalar adjectives pertain to various domain scales and vary in intensity within each scale (e.g. certain is more intense than likely on the likelihood scale). Scalar implicatures arise from the consideration of alternative statements which could have been made. They can be triggered by scalar adjectives and require listeners to reason pragmatically about them. Some scalar adjectives are more likely to trigger scalar implicatures than others. This phenomenon is referred to as scalar diversity. In this study, we probe different families of Large Language Models such as GPT-4 for their knowledge of the lexical semantics of scalar adjectives and one specific aspect of their pragmatics, namely scalar diversity. We find that they encode rich lexical-semantic information about scalar adjectives. However, the rich lexical-semantic knowledge does not entail a good understanding of scalar diversity. We also compare current models of different sizes and complexities and find that larger models are not always better. Finally, we explain our probing results by leveraging linguistic intuitions and model training objectives.

V. G. Shvydko

article addresses the experience of Fumio Kishida’s first year as Prime Minister of the Japanese government with respect to its economic policy vision and implementation. This policy is analyzed by comparing it to the policy pursued over the past ten years by PM Kishida’s predecessors in the office, on the one hand, and to the commitments he announced during the 2021 election campaign, on the other hand. The paper notes that, in its basic moments, the economic policy of the cabinet of the new leader of the ruling Liberal Democratic Party continues the course previously maintained by the team of then-PM Abe Shinzō, known as Abenomics. This is manifested in the specific use of basic instruments of macroeconomic policy, primarily in the monetary and tax areas, as well as in setting priorities for the policy to stimulate consumption and economic activity. Continuity of the course can be particularly illustrated by the soft monetary policy; a positive view of the depreciation of the Japanese national currency; the moderately expansionary fiscal policy relying on domestic borrowing; preference given to interests of the national corporate sector. Legacy inherited from the previous administrations also includes government measures to revitalize deferred private demand, promoting investment in R&D, venture, and innovative enterprises with a particular focus on regional economies. Kishida’s particular emphasis on invigorating redistributive mechanisms and increasing the share of wages in the total national income has not yet resulted in specific decisions and actions by the government. Kishida’s commitment to expand the number of beneficiaries of the capitalist market system as part of his idea of “new capitalism” is yet to be carried out.In recent months, the attention of the government and its economic team has largely been focused on overcoming the consequences of the disruption of transnational production, trade and logistics chains caused by the coronavirus pandemic and rising geopolitical tensions. In addressing this issue, the government prioritizes subsidizing businesses and households to partially offset energy and food price hikes, securing diversification or localization of critical links of trade and production chains, as well as the exclusion from them of politically undesirable or unstable locations. At this stage, however, the actions of the government are limited to setting relevant goals and plans to provide finance for programs with effectiveness yet to be proven.

Paksiutov Georgii D.

The concept of economic security has received special attention from the Japanese government since 2019: several specialized state agencies have been created, and in May 2022 the Diet approved the Act on Provision of Economic Security. Economic security-related notions have also been included in the new National Security Strategy, published in December 2022. In this document, economic security is defined as protection of ‘Japan’s national interests, such as peace, security, and economic prosperity, by carrying out economic measures.’ It shall be noted that the current understanding of economic security is different from the one previously assumed, which was largely concerned with the provision of natural resources and food. We observe the concrete measures proposed in connection with the concept of economic security and highlight that the new understanding of the concept has been introduced in relation to the ongoing trade war between the U.S. and China, in which Japan supports the U.S. in its efforts against the certain advanced-tech companies from China. From the point of view of the securitization theory, claims that certain industries of the PRC’s economy present a threat to Japan’s security could be seen as ‘speech-acts’ aiming to legitimize the employment of extraordinary political measures. We emphasize that the apparent success of such speech-acts is grounded in the Japanese experts’ assessment that there are emergent technologies with the possibility of dual commercial and military use. In conclusion, we examine the securitization of China’s economy by the Japanese government in its current regional and global context and argue that it could be interpreted both as a signal to the U.S. and as an attempt to gain leverage in the competition between the Japanese and the Chinese infrastructural projects in the Asia-Pacific

Simon Stepputtis, Joseph Campbell, Yaqi Xie et al.

Deception and persuasion play a critical role in long-horizon dialogues between multiple parties, especially when the interests, goals, and motivations of the participants are not aligned. Such complex tasks pose challenges for current Large Language Models (LLM) as deception and persuasion can easily mislead them, especially in long-horizon multi-party dialogues. To this end, we explore the game of Avalon: The Resistance, a social deduction game in which players must determine each other's hidden identities to complete their team's objective. We introduce an online testbed and a dataset containing 20 carefully collected and labeled games among human players that exhibit long-horizon deception in a cooperative-competitive setting. We discuss the capabilities of LLMs to utilize deceptive long-horizon conversations between six human players to determine each player's goal and motivation. Particularly, we discuss the multimodal integration of the chat between the players and the game's state that grounds the conversation, providing further insights into the true player identities. We find that even current state-of-the-art LLMs do not reach human performance, making our dataset a compelling benchmark to investigate the decision-making and language-processing capabilities of LLMs. Our dataset and online testbed can be found at our project website: https://sstepput.github.io/Avalon-NLU/

Lucas Fernando Souza de Castro, Sandro Rigo

Context: Microservices running and being powered by Edge Computing have been gaining much attention in the industry and academia. Since 2014, when Martin Fowler popularized the Microservice term, many studies have been published relating these subjects to explore how the Edge's low-latency feature could be combined with the high throughput of the distributed paradigm from Microservices. Objective: Identifying how Microservices work together with Edge Computing whereas they take advantage when running on Edge. Method: In order to better understand this relationship, we first identified its key concepts, which are: architecture approaches and features, microservice composition (orchestration/choreography), and offloading. Afterward, we conducted a Systematic Literature Review (SLR) as the survey method. Results: We reviewed 111 selected studies and built a taxonomy of Microservices on Edge Computing demonstrating their current architecture approaches and features, composition, and offloading modes. Moreover, we identify the research gaps and trends. Conclusion: This paper is a step forward to help researchers and professionals get a general overview of how Microservices and Edge have been related in the last years. It also discusses gaps and research trends. This SLR will also be a good introduction for new researchers in Edge and Microservices.

Sangwon Hyun, Mingyu Guo, M. Ali Babar

Large-Language Models (LLMs) have shifted the paradigm of natural language data processing. However, their black-boxed and probabilistic characteristics can lead to potential risks in the quality of outputs in diverse LLM applications. Recent studies have tested Quality Attributes (QAs), such as robustness or fairness, of LLMs by generating adversarial input texts. However, existing studies have limited their coverage of QAs and tasks in LLMs and are difficult to extend. Additionally, these studies have only used one evaluation metric, Attack Success Rate (ASR), to assess the effectiveness of their approaches. We propose a MEtamorphic Testing for Analyzing LLMs (METAL) framework to address these issues by applying Metamorphic Testing (MT) techniques. This approach facilitates the systematic testing of LLM qualities by defining Metamorphic Relations (MRs), which serve as modularized evaluation metrics. The METAL framework can automatically generate hundreds of MRs from templates that cover various QAs and tasks. In addition, we introduced novel metrics that integrate the ASR method into the semantic qualities of text to assess the effectiveness of MRs accurately. Through the experiments conducted with three prominent LLMs, we have confirmed that the METAL framework effectively evaluates essential QAs on primary LLM tasks and reveals the quality risks in LLMs. Moreover, the newly proposed metrics can guide the optimal MRs for testing each task and suggest the most effective method for generating MRs.

Nastasya Kirana Dewi, Raden Novitasari

Language is a means of communication and interaction among humans to share information from all over the world, therefore language has a very important role in human life. Slang is a variety of non-formal language used by people aged 10-30 years. Slang can also be used as a secret form by using words that can be understood by certain circles and are temporary. By using this slang, it means that we are people who follow the times. Slang can also be modified and can create new vocabulary. The purpose of this study is to examine and describe the form, meaning, and process of forming Japanese slang words found on the Nihongo Mantappu Jerome Polin youtube channel. The method of data collection used the method of observing with note taking techniques, and the method used in the analysis was a qualitative descriptive method. The research data is in the form of slang that appears on the Nihongo Mantappu Jerome Polin Youtube channel. The data collected were 19 Japanese slang data with various variations of forms and meanings. The theory used to analyze is the theory proposed by Kageyama (2016) regarding the process of word formation. The results obtained according to Kageyama's theory (2016) are the process of forming Japanese slang words consisting of compounding (C), affixation (A), conversion (K), reduplication (R), blending (B), clipping (C), and acronyms. or initials (I).

Prasetya Maulana Yasin, Linna Meilia Rasiban, Aep Saeful Bachri

Kandoushi is a word that expresses an impression, also called “interjection” in Indonesian, and is one type of emotive interjection. This study examines the meaning and use of emotive interjection, and its similarities and differences in Japanese and Indonesian. The method in this study used a descriptive contrastive analysis method with the note-taking technique. The data is taken from the dialogue in the Japanese anime “ReLIFE” and the Indonesian TV series “Tetangga Masa Gitu” and “Bajaj Bajuri”. The results of the study showed that the emotive interjection in Japanese and Indonesian had several similar meanings, particularly expressing feelings of pleasure, admiration, annoyance, confusion, and distress. Then, both of them have the same use, which is used to confirm the speech partner’s expression, is used to conclude something, is used to protest against the speech partner, is used to show approval, and is used to realize something. Interjection comprehension is very important in understanding an utterance intent in a conversation so that communication can take place with native speakers or fellow foreign students. This study only examines the interjections in terms of the emotive feeling/impression of being surprised. Therefore, for future research, it is expected to research form, meaning, and the use of other interjections in terms of phonology.

Suwako Uehara, Hibiya Haraki, Stuart McLean

Japanese graduate school students in the field of science and engineering need to read academic research in their second language (L2), and such tasks can be challenging. Studies showed a strong (0.78) correlation between vocabulary size and reading comprehension (McLean et al., 2020), and providing high-frequency word lists could enhance comprehension. In this work-in-progress, 1.35 million tokens of professor-recommended reading materials were used to investigate a method to create a vocabulary list that would benefit science majors in graduate school, the procedures to create a corpus and a high-frequency word list efficiently, and the steps required to create a cleaner corpus. This paper outlines a systematic literature-informed method that includes input from professors in the field, the combined use of tailored script in MATLAB and AntCont (Anthony, 2022) generated corpus and high-frequency words efficiently, and repeated comparison of original PDFs and the matching text files, then adding MATLAB script to deal with specific issues created by a cleaner text. This proposed method can be applied in other contexts to enhance the generation of high-frequency word lists.

Dimitrios Alivanistos, Selene Báez Santamaría, Michael Cochez et al.

Language Models (LMs) have proven to be useful in various downstream applications, such as summarisation, translation, question answering and text classification. LMs are becoming increasingly important tools in Artificial Intelligence, because of the vast quantity of information they can store. In this work, we present ProP (Prompting as Probing), which utilizes GPT-3, a large Language Model originally proposed by OpenAI in 2020, to perform the task of Knowledge Base Construction (KBC). ProP implements a multi-step approach that combines a variety of prompting techniques to achieve this. Our results show that manual prompt curation is essential, that the LM must be encouraged to give answer sets of variable lengths, in particular including empty answer sets, that true/false questions are a useful device to increase precision on suggestions generated by the LM, that the size of the LM is a crucial factor, and that a dictionary of entity aliases improves the LM score. Our evaluation study indicates that these proposed techniques can substantially enhance the quality of the final predictions: ProP won track 2 of the LM-KBC competition, outperforming the baseline by 36.4 percentage points. Our implementation is available on https://github.com/HEmile/iswc-challenge.

Rahayu Siwi Winarni, Linna Meilia Rasiban

The purpose of this research is to examine students' perceptions towards the use of YouTube as a teaching medium to increase students’ interest and motivation in learning Japanese. The method used is descriptive qualitative with survey techniques and literature research. The descriptive analysis is used to describe the percentage of each variable, namely the use of YouTube and Japanese students' interest and learning motivation. The results of this study indicate that YouTube offers a significant effect on students in the use of online video as a learning medium and the research findings show that the participants positively view the use of YouTube in their lessons. The result also revealed a significant increase in increasing student interest and motivation in learning Japanese. The results of this study can be an illustration to encourage students, especially in learning Japanese, to make the most of the online media or video channels on YouTube related to Japanese language learning. This research also can be an online learning system in technological development.

Joshua S. Mostow

The Ise monogatari (The Ise Stories, aka Tales of Ise, tenth cen.) is the oldest continuously illustrated secular narrative in Japanese history. The present article explores to what extent, and how, contemporary manga artists engage with or use this rich visual tradition, examining three examples, in the seinen (young male-oriented), shōjo (young female-oriented), and gyagu (gag) genres, yet all arguably categorizable as gakushū, or educational, manga. Perhaps surprisingly, only the gag manga artist, Kurogane Hiroshi, takes advantage of the Ise’s long visual history, and the author of the article concludes by drawing parallels with the early modern artistic practice of mitate-e, or visual parody.

Stefan Hoffmann

The projected language of a general deterministic automaton with $n$ states is recognizable by a deterministic automaton with $2^{n-1} + 2^{n-m} - 1$ states, where $m$ denotes the number of states incident to unobservable non-loop transitions, and this bound is best possible. Here, we derive the tight bound $2^{n - \lceil \frac{m}{2} \rceil} - 1$ for permutation automata. For a state-partition automaton with $n$ states (also called automata with the observer property) the projected language is recognizable with $n$ states. Up to now, these, and finite languages projected onto unary languages, were the only classes of automata known to possess this property. We show that this is also true for commutative automata and we find commutative automata that are not state-partition automata.