ACM Transactions on Asian and Low-Resource Language Information Processing

Journal Information
ISSN / EISSN : 2375-4699 / 2375-4702
Total articles ≅ 360
Current Coverage

Latest articles in this journal

Jun Ma, Hongzhi Yu, Yan Xu, Kaiying Deng
ACM Transactions on Asian and Low-Resource Language Information Processing, Volume 20, pp 1-10;

According to relevant specifications, this article divides, marks, and extracts the acquired speech signals of the Salar language, and establishes the speech acoustic parameter database of the Salar language. Then, the vowels of the Salar language are analyzed and studied by using the parameter database. The vowel bitmap (average value at the beginning of words), the vowel bitmap (average value at the abdomen of words), the vowel bitmap (average value at the ending of words), and the vowel bitmap (average value) are obtained. Through the vowel bitmaps, we can observe the vowel in different positions of the word, the overall appearance of an obtuse triangle. The high vowel [i], [o], and low vowel [a] occupy three vertices, respectively. Among the three lines, [i] to [o] are the longest, [i] to [a] are the second longest, and [a] to [o] are the shortest. The lines between [a] to [o] and [a] and [i] are asymmetric. Combining with the vowel bitmap, the vowels were discretized, and the second formant (F2) frequency parameter was used as the coordinate of the X axis, and the first formant (F1) frequency was used as the coordinate of the Y axis to draw the region where the vowel was located, and then the vowel pattern was formed. These studies provide basic data and parameters for the future development of modern phonetics such as the database of Sarah language speech, speech recognition, and speech synthesis. It also provides the basic parameters of speech acoustics for the rare minority acoustic research work of the national language project.
Weipeng Jing, Xianyang Song, Donglin Di, Houbing Song
ACM Transactions on Asian and Low-Resource Language Information Processing, Volume 20, pp 1-18;

In the area of geographic information processing, there are few researches on geographic text classification. However, the application of this task in Chinese is relatively rare. In our work, we intend to implement a method to extract text containing geographical entities from a large number of network texts. The geographic information in these texts is of great practical significance to transportation, urban and rural planning, disaster relief, and other fields. We use the method of graph convolutional neural network with attention mechanism to achieve this function. Graph attention networks (GAT) is an improvement of graph convolutional neural networks (GCN). Compared with GCN, the advantage of GAT is that the attention mechanism is proposed to weight the sum of the characteristics of adjacent vertices. In addition, We construct a Chinese dataset containing geographical classification from multiple datasets of Chinese text classification. The Macro-F Score of the geoGAT we used reached 95% on the new Chinese dataset.
Gunasekaran Manogaran, Hassan Qudrat-Ullah, Qin Xin
ACM Transactions on Asian and Low-Resource Language Information Processing, Volume 20, pp 1-3;

Chunhe Zhao, BalaAnand Muthu, P. Mohamed Shakeel
ACM Transactions on Asian and Low-Resource Language Information Processing, Volume 20, pp 1-16;

This research proposes to evaluate and analyze the decision matrix for learner's English mobile applications (EMAs) based on multi-objective heuristic decision making with a view to listening, speaking, reading, and writing. Because of the number of criteria, the significance of parameters, and variance in results, EMAs are difficult. Decision making has built on the combination of listening, speaking, reading, and writing and EMA evaluation criteria for students. The requirements are adapted from a framework of pre-school education. Six alternatives and 17 skills as a requirement are included in decision-making results. The six EMA are then assessed, with six English learning experts distributing a review form. The application subsequently is evaluated using the best-worst method and preference-order technique (TOPSIS) using multi-objective heuristic decision making methods. The best-worst method is used to measure requirements, whereas TOPSIS is used to test and assess the applications. In two cases, namely person and group, TOPSIS is used. Internal and external aggregations are used throughout the group context. In effect, the aim of evaluating the proposed study and comparing it to six relative studies with scenarios and benchmarking checklists is to develop an objectives validation framework for e-apps.
Meng Li
ACM Transactions on Asian and Low-Resource Language Information Processing, Volume 20, pp 1-16;

To effectively identify the influencing factors of the perceived usefulness of multimodal data in online reviews of tourism products, this article explores the optimization method of online tourism products based on user-generated content and conducts feature fusion of multimodal data in online reviews of tourism products from the perspective of data fusion analysis. Therefore, based on the word vector model, this article proposes a method to select the seed word set of emotion dictionary. In this method, emotional words are represented in vector form and the distance between word vectors is calculated to form the selection criteria and classification basis of seed word set, and then the sentiment dictionary of online review is formed by category judgment. This article takes the real online review data of tourism products as the research object, carries out descriptive statistical analysis, uses machine learning and deep learning methods, carries out text vector embedding and image content recognition, integrates image and text feature vector, constructs multimodal online review usefulness classification model, and conducts model test. The experimental results show that, compared with the single-mode reviews containing only text or pictures, the multimodal reviews combined with text and pictures can better predict the usefulness of online reviews, improve the quality of online reviews, give full play to the potential value of user-generated content, provide optimization ideas for product providers, and provide decision support for product consumers.
Karunesh Kumar Arora, Shyam Sunder Agrawal
ACM Transactions on Asian and Low-Resource Language Information Processing, Volume 20, pp 1-18;

English and Hindi have significantly different word orders. English follows the subject-verb-object (SVO) order, while Hindi primarily follows the subject-object-verb (SOV) order. This difference poses challenges to modeling this pair of languages for translation. In phrase-based translation systems, word reordering is governed by the language model, the phrase table, and reordering models. Reordering in such systems is generally achieved during decoding by transposing words within a defined window. These systems can handle local reorderings, and while some phrase-level reorderings are carried out during the formation of phrases, they are weak in learning long-distance reorderings. To overcome this weakness, researchers have used reordering as a step in pre-processing to render the reordered source sentence closer to the target language in terms of word order. Such approaches focus on using parts-of-speech (POS) tag sequences and reordering the syntax tree by using grammatical rules, or through head finalization. This study shows that mere head finalization is not sufficient for the reordering of sentences in the English-Hindi language pair. It describes various grammatical constructs and presents a comparative evaluation of reorderings with the original and the head-finalized representations. The impact of the reordering on the quality of translation is measured through the BLEU score in phrase-based statistical systems and neural machine translation systems. A significant gain in BLEU score was noted for reorderings in different grammatical constructs.
, Sukomal Pal, Chiranjeev Kumar
ACM Transactions on Asian and Low-Resource Language Information Processing, Volume 20, pp 1-34;

With Web 2.0, there has been exponential growth in the number of Web users and the volume of Web content. Most of these users are not only consumers of the information but also generators of it. People express themselves here in colloquial languages, but using Roman script (transliteration). These texts are mostly informal and casual, and therefore seldom follow grammar rules. Also, there does not exist any prescribed set of spelling rules in transliterated text. This freedom leads to large-scale spelling variations, which is a major challenge in mixed script information processing. This article studies different existing phonetic algorithms to handle the issue of spelling variation, points out the limitations of them, and proposes a novel phonetic encoding approach with two different flavors in the light of Hindi transliteration. Experiments performed over Hindi song lyrics retrieval in mixed script domain with three different retrieval models show that proposed approaches outperform the existing techniques in a majority of the cases (sometimes statistically significantly) for a number of metrics like [email protected], [email protected], [email protected], MAP, MRR, and Recall.
R. Lavanya, B. Bharathi
ACM Transactions on Asian and Low-Resource Language Information Processing, Volume 20, pp 1-14;

With the increase in numbers of multimedia technologies around us, movies and videos on social media and OTT platforms are growing, making it confusing for users to decide which one to watch for. For this, movie recommendation systems are widely used. It has been observed that two-thirds of the films watched on Netflix are the recommended ones to its users. The target of this work is to use implicit feedback given by other users to recommend movies, i.e., ratings given by them. Implicit feedback will help to enhance Data Sparsity as for a replacement logged-in user, the system won't have details of their past liked movies. So, matching the similarity with other users is often a plus point to recommend movies that they would like. The anticipated result will depend upon the positive attitude; i.e., if the predicted rating is high, then it'll be recommended; otherwise it'll not be recommended. The performance of the methodology is measured with accuracy and precision values for different strategies. It gives the best accuracy and highest precision values using Logistic Regression (LR) and lowest recall value as compared to other algorithms. This technique gives an accuracy, precision, and recall value of 81.9%, 69.82%, and 32.5%, respectively, using LR.
Lin Sun, Wenzheng Xu, Jimin Liu
ACM Transactions on Asian and Low-Resource Language Information Processing, Volume 20, pp 1-12;

Using hierarchical CNN, the company's multiple news is characterized as three levels: sentence vectors, chapter vectors, and enterprise sentiment vectors. By combining the stock price data with the news lyric data at the same time, the influence of news on price is used to achieve correlation analysis of news information and stock prices. A two-channel attention mechanism fusion model based on CNN-LSTM is proposed. After the dual-channel feature extraction, the attention layer fusion layer is used to convert the weighted values of LSTM hidden variables, so the stock price can be predicted with the news text.
Junyang Tan, Dan Xia, Shiyun Dong, Honghao Zhu, Binshi Xu
ACM Transactions on Asian and Low-Resource Language Information Processing, Volume 20, pp 1-15;

The Internet of Things and big data are currently hot concepts and research fields. The mining, classification, and recognition of big data in the Internet of Things system are the key links that are widely of concern at present. The artificial neural network is beneficial for multi-dimensional data classification and recognition because of its strong feature extraction and self-learning ability. Pre-training is an effective method to address the gradient diffusion problem in deep neural networks and could result in better generalization. This article focuses on the performance of supervised pre-training that uses labelled data. In particular, this pre-training procedure is a simulation that shows the changes in judgment patterns as they progress from primary to mature within the human brain. In this article, the state-of-the-art of neural network pre-training is reviewed. Then, the principles of the auto-encoder and supervised pre-training are introduced in detail. Furthermore, an extended structure of supervised pre-training is proposed. A set of experiments are carried out to compare the performances of different pre-training methods. These experiments include a comparison between the original and pre-trained networks as well as a comparison between the networks with two types of sub-network structures. In addition, a homemade database is established to analyze the influence of pre-training on the generalization ability of neural networks. Finally, an ordinary convolutional neural network is used to verify the applicability of supervised pre-training.
Back to Top Top