Article
Education & Educational Research
Mutahar Qassem
Summary: This study examines the cognitive processes of Arab translator trainees when translating English news articles into Arabic, and finds that revision time, orientation time, translation speed, and translation units are related to translation adequacy.
INTERACTIVE LEARNING ENVIRONMENTS
(2022)
Article
Telecommunications
Yumin Shen, Hongyu Guo
Summary: Retelling extraction is a crucial aspect of NLP, and having high-quality retelling resources is beneficial for enhancing machine translation. However, traditional methods often neglect the document background during retelling acquisition and application. To address this issue, we incorporate topic model information into the translation process and propose a topic-based statistical machine translation method for improving translation performance. Our method utilizes Probabilistic Latent Semantic Analysis (PLSA) to establish the co-occurrence relationship between words and documents through hybrid matrix decomposition. Furthermore, we introduce a decoder to simplify the decoding process. Experimental results demonstrate the effectiveness of the proposed method in enhancing translation accuracy.
DIGITAL COMMUNICATIONS AND NETWORKS
(2023)
Article
Computer Science, Information Systems
Sajadul Hassan Kumhar, Syed Immamul Ansarullah, Akber Abid Gardezi, Shafiq Ahmad, Abdelaty Edrees Sayed, Muhammad Shafiq
Summary: English to Urdu machine translation is still in its early stages and lacks effective translation methods. Machine translation has made knowledge accessible to the masses by automating translation from source language to target language. Various approaches, including rule-based, computational, hybrid, and neural machine translation, have been proposed. This research employs a neural machine translation approach using LSTM Encoder-Decoder to translate English text into Urdu. The results show improved performance, evaluated through bilingual research metrics, with the highest score sequences having an effective length of 10.
CMC-COMPUTERS MATERIALS & CONTINUA
(2023)
Article
Multidisciplinary Sciences
Daniel Vujcich, Meagan Roberts, Zhihong Gu, Shih-Chi Kao, Roanna Lobo, Limin Mao, Enaam Oudih, Nang Nge Nge Phoo, Horas Wong, Alison Reid
Summary: Survey translation is time-consuming and costly, leading to deviations from TRAPD 'best practice'. However, our study demonstrates that even adaptations of the TRAPD method can identify issues that may not have been apparent with non-team-based or single-round translation approaches.
Article
Mathematics
Yuexiang Ruan
Summary: The intercommunication and translation of global languages have become crucial for friendly communication among people worldwide. This study improved machine translation using deep learning networks and focused on solving the out-of-vocabulary problem and incomplete translation. Experimental results showed that the proposed methods effectively addressed these issues and improved translation accuracy.
JOURNAL OF MATHEMATICS
(2022)
Article
Psychology, Multidisciplinary
Yixiao Cui, Binghan Zheng
Summary: This study compares linguistic and extralinguistic consultations in English-Chinese translation in terms of cognitive resources allocation and information processing patterns using eye-tracking and screen-recording data. It also explores the differences among different types of translators and the effect of extralinguistic consultation on translation quality.
FRONTIERS IN PSYCHOLOGY
(2022)
Article
Multidisciplinary Sciences
Mohammad Ahmad Thawabteh, Amer Al-Adwan, Amna Shqair
Summary: This article examines the translation of taboo language in the Lebanese movie "West Beirut" and explores the challenges and strategies in subtitling. The study finds that subtitlers employ various translation strategies to deal with excessive taboo language, some of which may distort the original dialogue while others effectively reach the intended audience. Cultural considerations play a significant role in determining the choice of translation strategies and their frequency.
Review
Social Sciences, Interdisciplinary
Kate Benedicta Amenador, Zhiwei Wang
Summary: The research on Chinese-English public sign translation in China has gone through different stages of development, with fruitful results achieved. However, there are still some deficiencies, such as repetition of research content and limited scope of investigation. Future research should be more diversified and in-depth, and the coverage of areas of research should be expanded.
Article
Computer Science, Artificial Intelligence
Salam Michael Singh, Thoudam Doren Singh
Summary: This study presents a semi-supervised neural machine translation system that improves translation quality for low-resource language pairs. By introducing external noise and utilizing a multi-reference test dataset, the proposed system outperforms other methods significantly.
EXPERT SYSTEMS WITH APPLICATIONS
(2022)
Article
Integrative & Complementary Medicine
Lily Lim, Ping Chen, Vincent Xian Wang
Summary: Translating TCM terminology is difficult but essential for medical practices outside China. There are diverse and inconsistent translations in bilingual glossaries and dictionaries, highlighting the need to study the strengths and weaknesses of different translations. The study emphasizes the importance of creating a comprehensive bilingual reference work containing representative translations to support clinical practice and the spread of TCM.
EUROPEAN JOURNAL OF INTEGRATIVE MEDICINE
(2022)
Article
Computer Science, Information Systems
Xin Ke
Summary: Machine translation has been a crucial and challenging topic in natural language processing. This work proposes and implements a real-time synchronous translation method, focusing on solving key technology issues in translation generation. By improving template selection and phrase translation, the effectiveness of machine translation is enhanced.
Article
Psychology, Mathematical
Soon Tat Lee, Walter J. B. van Heuven, Jessica M. Price, Christine Xiang Ru Leong
Summary: This study presents the first Malay-English translation norms collected with highly proficient Malay-English bilinguals and investigates the impact of lexical characteristics on translation ambiguity. The results show that verbs, adjectives, and class-ambiguous words are more translation-ambiguous, and within-language semantic variability and word length are positively correlated with translation ambiguity.
BEHAVIOR RESEARCH METHODS
(2023)
Article
Psychology, Multidisciplinary
Huiping Wang
Summary: This study reveals the translator's voice in six English versions of the Chinese classic Daxue by analyzing language and non-language choices made by each translator. The findings suggest that the translator's cultural identity, historical background, and motives are evident in the translations. The study highlights the correlation between the translator's voice and the cultural and ideological constructs of translation.
FRONTIERS IN PSYCHOLOGY
(2023)
Article
Multidisciplinary Sciences
Mohammad Ahmad Thawabteh, Amer Al-Adwan
Summary: This study focuses on the voice-over subfield of audiovisual translation, specifically the English-Arabic translation of a BBC documentary. It finds that student translators encounter technical and linguistic challenges, as well as potential semantic and stylistic loss, emphasizing the pedagogical angle for trainers and trainees in the voice-over mode.
Article
Computer Science, Hardware & Architecture
Tianyuan Huang, Wei Su, Lei Liu, Jing Zhang, Chuan Cai, Yongna Yuan, Cunlu Xu
Summary: Transforming Mandarin Braille to Chinese text is a less focused machine translation task. CBHG, a building block in the Tacotron text-to-speech model, can be used to perform Braille-Chinese translation. However, relying solely on the convolution blocks in CBHG is not effective in extracting Braille sequence features. Two approaches, CBHG-SE and CBHG-ECA, are proposed to improve the CBHG model by adaptively recalibrating channel-wise feature responses and utilizing global information. CBHG-ECA shows stronger feature recalibration capabilities and both models outperform other neural machine translation models with BLEU scores of 92.23 and 91.48 on the Braille-Chinese dataset.