Search Results (23)

This study proposes a new benchmark to evaluate the cultural understanding and natural language processing capabilities of large language models based on Sino-Korean words and four-character idioms. Those are essential linguistic and cultural assets in Korea. Reflecting the official question types of the Korean Hanja Proficiency Test, we constructed four question categories—four-character idioms, synonyms, antonyms, and homophones—and systematically compared the performance of GPT-based and non-GPT LLMs. GPT-4o showed the highest accuracy and explanation quality. However, challenges remain in distinguishing the subtle nuances of individual characters and in adapting to uniquely Korean meanings as opposed to standard Chinese character interpretations. Our findings reveal a gap in LLMs’ understanding of Korea-specific Hanja culture and underscore the need for evaluation tools reflecting these cultural distinctions. Full article

Chinese, a tonal language with inherent homophonic ambiguity, poses significant challenges for semantic disambiguation in natural language processing (NLP), hindering applications like speech recognition, dialog systems, and assistive technologies. Traditional static disambiguation methods suffer from poor adaptability in dynamic environments and low-frequency scenarios, limiting their real-world utility. To address these limitations, we propose BLAF—a novel MacBERT-BiLSTM Hybrid Architecture—that synergizes global semantic understanding with local sequential dependencies through dynamic multimodal feature fusion. This framework incorporates innovative mechanisms for the principled weighting of heterogeneous features, effective alignment of representations, and sensor-augmented cross-modal learning to enhance robustness, particularly in noisy environments. Employing a staged optimization strategy, BLAF achieves state-of-the-art performance on the SIGHAN 2015 (data fine-tuning and supplementation): 93.37% accuracy and 93.25% F1 score, surpassing pure BERT by 15.74% in accuracy. Ablation studies confirm the critical contributions of the integrated components. Furthermore, the sensor-augmented module significantly improves robustness under noise (speech SNR to 18.6 dB at 75 dB noise, 12.7% reduction in word error rates). By bridging gaps among tonal phonetics, contextual semantics, and computational efficiency, BLAF establishes a scalable paradigm for robust Chinese homophone disambiguation in industrial NLP applications. This work advances cognitive intelligence in Chinese NLP and provides a blueprint for adaptive disambiguation in resource-constrained and dynamic scenarios. Full article

Financial text analytics methods are employed to examine social media comments, allowing investors to gain insights and make informed financial decisions. Some emojis within these comments often convey diverse semantics, emotions, or intentions depending on the context. However, traditional financial text analysis methods relying on public annotations struggle to identify implicit expressions, leading to suboptimal performance. To address this challenge, this paper proposes an implicit expression recognition model of emojis in social media comments (IER-SMCEM). Firstly, IER-SMCEM innovative designs a data enhancement method based on the implicit expression of emoji. This method expands the pure text financial sentiment analysis dataset into the implicit expression dataset of emoji by homophonic replacement. Secondly, IER-SMCEM designs a prompt learning template to identify the implicit expression of emoji. Through hand-designed templates, large-scale language models can predict the true meaning that emojis are most likely to express. Finally, IER-SMCEM recovers implicit expression by choosing the predictions of models. Thus, the downstream financial sentiment analysis model can more precisely realize the sentiment recognition of the text with emoji by the recovered text. The experimental results indicate that IER-SMCEM achieves a 98.03% accuracy in semantically recovering implicit expressions within financial texts. In the task of financial sentiment analysis, the sentiment analysis model achieves the highest accuracy of 3.99% after restoring the true implied expression of the texts. Therefore, the model can be effectively applied to sentiment analysis or quantitative analysis. Full article

During conversation, speakers produce reduced speech, and this can create homophones: ‘we were’ and ‘we’re’ can both be realized as [ɚ], and ‘he was’ and ‘he’s’ can be realized as [ɨz]. We investigate the types of information non-native listeners (Dutch L1-English L2) use to perceive the tense of such verbs, making comparisons with previous results from native listeners. The Dutch listeners were almost as successful as natives (average percentage correct for ‘is’/’was’ in the most accurate condition: 81% for Dutch, 88% for natives). The two groups showed many of the same patterns, indicating that both make strong use of whatever acoustic information is available in the signal, even if it is heavily reduced. The Dutch listeners showed one crucial difference: a minimal amount of context around the target, just enough to signal speech rate, did not help Dutch listeners to recover the longer forms, i.e., was/were, from reduced pronunciations. Only the full utterance context (containing syntactic/semantic information such as ‘yesterday’ or another tensed verb) helped Dutch listeners to recover from reduction. They were not able to adjust their criteria based on the surrounding speech rate as native listeners were. This study contributes to understanding how L2 learners parse information from spontaneous speech in a World Englishes setting with inputs from multiple dialects. Full article

Mo Yan’s 2009 novel Frog (Wā 蛙) traces the dramatic career of a rural obstetrician who saves lives through modern medicine, forces vasectomies and abortions through her implementation of the one-child policy, supports her nephew’s black market surrogacy scheme, and finally ends up withdrawing into a spiritual state of atonement for her previous deeds. This article examines the relationship between human and animal in the novel, suggesting that the conceptual separation of these categories is intimately related to the various problems the novel depicts throughout Chinese modernity. By focusing on the critical possibilities offered by the novel’s title, Wā 蛙, as a homophone with both “baby” (wá 娃) and the “wa” of the mythical female progenitor Nüwa (娲wā), I suggest that Mo Yan offers a new concept of life, best referred to simply as wa, in response to certain crises of modernity. As an ambiguously generative reconceptualization of life, wa denies conventional and simplistic distinctions between human and animal while incorporating elements of spirituality and unknowability into an otherwise overly rationalized and monetized idea of the human. Full article

Two experiments were conducted to examine native and non-native speakers’ recognition of Chinese two-character words (2C-words) in the context of audio sentence comprehension. The recording was played of a sentence, in which a collocation composed of a number word, a sortal classifier, and a noun (NCN) was embedded. When the participants were about to hear the noun of the NCN (Noun), the playing stopped, and a target was visually presented, which was the Noun, the character-transposed word of the Noun (NounT), or a control word (NounC), or was a homophone nonword for Noun, NounT, or NounC. The participants were required to make a lexical decision on the target before they resumed listening. The results showed that both native and non-native speakers were able to take visually presented 2C-word targets as semantic whole entities in the context of audio sentence comprehension, which was mediated by their Chinese proficiency. Native speakers readily processed visually presented 2C-words both as wholes and according to their constituent characters, but non-native speakers were not likely to process the 2C-words according to their constituent characters. Full article

The effect of cross-linguistic similarity on the development of target-like categories in a second or additional language is widely attested. Research also shows that second-language speakers may access both their native and the second-language lexicons when processing second-language speech. Forty-three Catalan learners of English performed a perceptual assimilation task evaluating the perceived similarity between English and Catalan vowels and also participated in a visual world eye-tracking experiment investigating between-language lexical competition. The focus of the study was the English vowel contrasts /iː/-/ɪ/ and /æ/-/ʌ/. The perceptual task confirmed that English /iː/ and /æ/were perceptually closer to native Catalan categories than English /ɪ/ and /ʌ/. The results of the spoken word recognition task indicated that learners experienced greater competition from native words when the target words contained English /iː/ and /æ/, illustrating a close link between the two types of tasks. However, differences in the magnitude of cross-language lexical competition were found to be only weakly related to learners’ degree of perceived similarity to native categories at an individual level. We conclude that online tasks provide a potentially effective method of assessing cross-linguistic similarity without the concerns inherent to more traditional offline approaches. Full article

The topic of the article is the status of translation and homophony in philosophy, psychoanalysis and philology. The article focuses on the question of how translation is carried out using the basic principle of equivalence of meaning by homophony and what effects this can produce. The analysis of two case studies by Freud and Lacan shows that homophonic transfer from one language to another can be extremely productive for the subjective traversal of a phantasm. It is then shown that this is not, however, of purely subjective interest. Werner Hamacher has sketched the future of philology starting from such homophonic translations; Lacan has tried to advance to another theory of language through homophonic formations. Full article

Along with the explosion of ChatGPT, the artificial intelligence question-answering system has been pushed to a climax. Intelligent question-answering enables computers to simulate people’s behavior habits of understanding a corpus through machine learning, so as to answer questions in professional fields. How to obtain more accurate answers to personalized questions in professional fields is the core content of intelligent question-answering research. As one of the key technologies of intelligent question-answering, the accuracy of text matching is related to the development of the intelligent question-answering community. Aiming to solve the problem of polysemy of text, the Enhanced Representation through Knowledge Integration (ERNIE) model is used to obtain the word vector representation of text, which makes up for the lack of prior knowledge in the traditional word vector representation model. Additionally, there are also problems of homophones and polyphones in Chinese, so this paper introduces the phonetic character sequence of the text to distinguish them. In addition, aiming at the problem that there are many proper nouns in the insurance field that are difficult to identify, after conventional part-of-speech tagging, proper nouns are distinguished by especially defining their parts of speech. After the above three types of text-based semantic feature extensions, this paper also uses the Bi-directional Long Short-Term Memory (BiLSTM) and TextCNN models to extract the global features and local features of the text, respectively. It can obtain the feature representation of the text more comprehensively. Thus, the text matching model integrating BiLSTM and TextCNN fusing Multi-Feature (namely MFBT) is proposed for the insurance question-answering community. The MFBT model aims to solve the problems that affect the answer selection in the insurance question-answering community, such as proper nouns, nonstandard sentences and sparse features. Taking the question-and-answer data of the insurance library as the sample, the MFBT text-matching model is compared and evaluated with other models. The experimental results show that the MFBT text-matching model has higher evaluation index values, including accuracy, recall and F1, than other models. The model trained by historical search data can better help users in the insurance question-and-answer community obtain the answers they need and improve their satisfaction. Full article

Technology-based in-home reading and spelling programs have the potential to compensate for the lack of sufficient instructions provided at schools. However, the recent COVID-19 pandemic showed the immaturity of the existing remote teaching solutions. Consequently, many students did not receive the necessary instructions. This paper presents a model for developing intelligent reading and spelling programs. The proposed approach is based on an optimization model that includes artificial neural networks and linear regression to maximize the educational value of the pedagogical content. This model is personalized, tailored to the learning ability level of each user. Regression models were developed for estimating the lexical difficulty in the literacy tasks of auditory and visual lexical decision, word naming, and spelling. For building these regression models, 55 variables were extracted from French lexical databases that were used with the data from lexical mega-studies. Forward stepwise analysis was conducted to identify the top 10 most important variables for each lexical task. The results showed that the accuracy of the models (based on root mean square error) reached 88.13% for auditory lexical decision, 89.79% for visual lexical decision, 80.53% for spelling, and 83.86% for word naming. The analysis of the results showed that word frequency was a key predictor for all the tasks. For spelling, the number of irregular phoneme-graphemes was an important predictor. The auditory word recognition depended heavily on the number of phonemes and homophones, while visual word recognition depended on the number of homographs and syllables. Finally, the word length and the consistency of initial grapheme-phonemes were important for predicting the word-naming reaction times. Full article

Dyslexia is a learning disorder related to receptive language characterized by difficulties with decoding, fluent word recognition, automatic naming skills and/or reading comprehension skills. It usually leads to severe functional impairment and the permanent need for support and interventions. Since animal-assisted interventions (AAIs) have been found to improve physical, emotional, cognitive and/or social functioning in humans, the aim of this study is to demonstrate the effectiveness of onotherapy on children with SLD by improving self-esteem and school performance. Sixteen patients with a diagnosis of dyslexia were randomly assigned to two treatment groups: the first was a conventional neuropsychological group therapy without onotherapy, and the second was a neuropsychological group therapy incorporating AAIs with therapy donkeys. The neuropsychological assessment included the WISC-IV, DDE and the TMA test, which were administered before and after the treatment in both groups. The results of the experimental group show significant improvement in word reading test correctness (p = 0.03) and speed (p = 0.03), non-word reading test speed (p = 0.01), reading text test correctness (p = 0.05) and speed (p = 0.03), word writing test correctness (p = 0.01), non-word writing test correctness (p = 0.02), writing sentences with homophonic words correctness (p = 0.01), interpersonal TMA (p = 0.04) and the total TMA (p = 0.04), which were significative. On the other hand, in the control group, significant differences were found in word reading test speed (p = 0.01), non-word reading test speed (p = 0.04), reading text test speed (p = 0.02), writing word test correctness (p = 0.01), writing non-word test correctness (p = 0.01) and writing sentences with homophonic words (p = 0.01). However, in this group, we observed no significant difference in the esteem of children. Training associated with the donkeys determined improved scholastic performances as far as reading is concerned and a change in self-esteem. Therefore, we can state that AAIs for dyslexia could be a viable and effective option to enhance the rehabilitation process, increase self-esteem and improve cognitive functions and language skills recovery. Full article

In recent years, the study of dyslexia has seen rapid progress in definition and classification, neuropsychological correlates, neurobiological factors, and intervention. However, there are few studies on how virtual reality can affect improving cognitive domains and cross-cutting pedagogical skills. We, therefore, tested intervention through the use of a virtual reality rehabilitation system (VRRS) in children with dyslexia. Twenty-eight patients diagnosed with dyslexia were enrolled in this study. One-half underwent conventional neuropsychological treatment, and the other half performed VR neurorehabilitation training using the VRRS. All patients were evaluated by neuropsychological assessment at baseline (T0) and at the end of the protocol (T1). The assessment included the administration of the Wechsler Intelligence Scale for Children-IV and the Italian Battery for the Assessment of Dyslexia and Dysorthography. Our results showed a significant difference in word-reading test scores as well as in homophonic writing. In addition, treatment type was found to affect some domains of the WISC. We believe that the VRRS led to improved outcomes through the use of VR, which encourages active exploration, improves engagement, and provides motivation and enjoyment, allowing longer training sessions and improving treatment adherence. Full article

Although much has been written in recent years on the emergence of non-binary English linguistic innovations, comparatively little has been written on non-binary French forms, especially neo-morphemes marking non-binary gender on nouns. As French is a grammatical-gender language with a traditionally binary (masculine and feminine) system, many non-binary Francophones have circumvented the social connection between grammatical gender and human gender in innovating new, non-binary markers for animate nouns and their modifiers. This study uses a mixed methods approach, combining analysis of non-binary French grammars alongside interview data in order to highlight the divergent morphological approaches underlying non-binary marking systems. Three approaches to the formation of non-binary nouns are identified: A Compounding Approach, which combines masculine and feminine markers; a Systematic Approach, which phonologically conditions the use of non-binary allomorphs, with the markers themselves ranging from phonologically novel within French syllabic structure, to homophonous with masculine and/or feminine variants; and an Invariable Approach, which applies a single non-binary marker across all nouns. Ultimately, this study disentangles both morphological patterns in the formation of non-binary words and some of the motivations behind them in an emerging French subtype well-known to be heterogeneous. Full article

The Malay World has been home to a range of social formations, from nomadic hunter-gatherers on land and sea, through (semi-)sedentary swiddeners and forest traders, to state-incorporated peasants and aristocrats. In their religious and secular musics, these populations display differing performance manners and organisation that reflect their distinctive socio-cultural and religious orientations. The musics serve to embed those orientations as aesthetically felt rather than conceptually talked about. The differences are encoded mainly onto contrasts between, on the one hand, highly heterophonic and/or starkly non-melismatic performance and, on the other, more homophonic and/or melismatic styles. Full article

The proliferation of spam in China has a negative impact on internet users’ experiences online. Existing methods for detecting spam are primarily based on machine learning. However, it has been discovered that these methods are susceptible to adversarial textual spam that has frequently been imperceptibly modified by spammers. Spammers continually modify their strategies to circumvent spam detection systems. Text with Chinese homophonic substitution may be easily understood by users according to its context. Currently, spammers widely use homophonic substitution to break down spam identification systems on the internet. To address these issues, we propose a Bidirectional Gated Recurrent Unit (BiGRU)–Text Convolutional Neural Network (TextCNN) hybrid model with joint embedding for detecting Chinese spam. Our model effectively uses phonetic information and combines the advantages of parameter sharing from TextCNN with long-term memory from BiGRU. The experimental results on real-world datasets show that our model resists homophone noise to some extent and outperforms mainstream deep learning models. We also demonstrate the generality of joint textual and phonetic embedding, which is applicable to other deep learning networks in Chinese spam detection tasks. Full article