Trường Đại học Công nghệ Thông tin, Đại học Quốc gia Thành phố Hồ Chí Minh
Công bố khoa học tiêu biểu
Sắp xếp:
Multi-stage transfer learning with BERTology-based language models for question answering system in vietnamese
International Journal of Machine Learning and Cybernetics - Tập 14 - Trang 1877-1902 - 2023
With the fast growth of information science and engineering, a large number of textual data generated are valuable for natural language processing and its applications. Particularly, finding correct answers to natural language questions or queries requires spending tremendous time and effort in human life. While using search engines to discover information, users manually determine the answer to a given question on a range of retrieved texts or documents. Question answering relies heavily on the capability to automatically comprehend questions in human language and extract meaningful answers from a single text. In recent years, such question–answering systems have become increasingly popular using machine reading comprehension techniques. On the other hand, high-resource languages (e.g., English and Chinese) have witnessed tremendous growth in question-answering methodologies based on various knowledge sources. Besides, powerful BERTology-based language models only encode texts with a limited length. The longer texts contain more distractor sentences that affect the QA system performance. Vietnamese has a variety of question words in the same question type. To address these challenges, we propose ViQAS, a new question–answering system with multi-stage transfer learning using language models based on BERTology for a low-resource language such as Vietnamese. Last but not least, our QA system is integrated with Vietnamese characteristics and transformer-based evidence extraction techniques into an effective contextualized language model-based QA system. As a result, our proposed system outperforms our forty retriever-reader QA configurations and seven state-of-the-art QA systems such as DrQA, BERTserini, BERTBM25, XLMRQA, ORQA, COBERT, and NeuralQA on three Vietnamese benchmark question answering datasets.
New sufficient criteria for epsilon-contraction of a class of nonlinear diference system with continuous time
Tạp chí Khoa học Tự nhiên Đại học Quốc gia Thành phố Hồ Chí Minh - Tập 3 Số 3 - Trang 213-224 - 2019
Contraction property of dynamical systems, especially difference systems, is one of the qualitative properties which have attracted much attention from many researchers for recent decades. Contraction of dynamical systems has many practical applications which means that two trajectories of the system convergence to each other when the time reaches to positive infinity. In this paper, by improving some existing approaches, we present a new approach to contraction problem of a class of nonlinear time-varying delay difference system with continuous time. We generalize the definition of contraction to -contraction. Then, we give some new explicit sufficient criteria for -contraction and global exponential stability of the mentioned system. Furthermore, we investigate-contraction of perturbed difference systems with continuous time under nonlinear perturbations in which perturbations are general time-varying functions. Then we obtain a new explicit-contraction bound for such systems subject to nonlinear time-varying perturbations. The obtained theorems generalize some existing results in the literature as particular cases. An example is given to illustrate the obtained results.
#contraction bound #contraction #exponential stability #perturbed systems #difference systems with continuous time
Tổng quan hệ thống về phát triển chương trình đào tạo năng lực thông tin phục vụ nghiên cứu khoa học cho sinh viên
Tạp chí Khoa học Đại học Đồng Tháp - - 2023
Nghiên cứu được thực hiện bằng phương pháp tổng quan hệ thống dựa trên dữ liệu các bài báo khoa học trong giai đoạn 2019 đến 29/8/2023. Kết quả nghiên cứu đã nêu lên các khía cạnh về phát triển chương trình đào tạo năng lực thông tin ở các cơ sở giáo dục đại học trên thế giới hiện nay bao gồm: nội dung đào tạo, hình thức đào tạo, phương pháp đào tạo, nhân sự đào tạo, quy mô đào tạo và phương pháp kiểm tra, đánh giá
#Chương trình đào tạo #năng lực thông tin #năng lực thông tin nghiên cứu khoa học #sinh viên #tổng quan hệ thống.
Dưới vi phân parabolic và áp dụng vào nghiên cứu điều kiện tối ưu
Trong bài báo này chúng tôi đề xuất khái niệm dưới vi phân parabolic thông qua dưới đạo hàm parabolic. Bên cạnh đó, chúng tôi trình bày một số tính chất của dưới vi phân parabolic cũng như các áp dụng của dưới vi phân parabolic vào nghiên cứu điều kiện tối ưu. Hơn nữa, trong bài báo này chúng tôi cũng xây dựng ví dụ minh họa cho các kết quả đạt được.
#Dưới đạo hàm parabolic #dưới vi phân parabolic #điều kiện tối ưu #nghiệm cô lập tĩnh địa phương
ViMRC - VLSP 2021: Context-Aware Answer Extraction in Vietnamese Question Answering
VNU Journal of Science: Computer Science and Communication Engineering - Tập 38 Số 2 - 2022
MRC is challenging the natural language processing fields; machines automatically have to answer questions based on specific passages for this task. In recent years, machine reading comprehension (MRC) has received much attention; many articles have been written about this task. However, most of those articles only develop models in two main languages, English and Chinese. In this article, we propose to apply a new model to the task of reading comprehension in Vietnamese. Specifically, we use BLANC (BLock AttentioN for Context prediction) on pre-trained baseline models to solve the Machine reading comprehension (MRC) task on Vietnamese. We have achieved good results when using BLANC on the baseline model. Specifically, with the MRC task at the VLSP share-task 2021, we scored 76.877% of F1-score on the private test and ranked 2nd in the total. This shows that BLANC works very well in MRC tasks and further enhances the Vietnamese MRC development.
USING SUM MATCH KERNEL WITH BALANCED LABEL TREE FOR LARGE-SCALE IMAGE CLASSIFICATION
Tạp chí tin học và điều khiển học - Tập 32 Số 2 - 2016
Large-scale image classification is a fundamental problem in computer vision due to many real applications in various domains. A label tree-based classification is one of effective approaches for reducing the testing complexity with a large number of class labels. However, how to build a label tree structure with cost efficiency and high accuracy classification is a challenge. The popular building tree method is to apply a clustering algorithm to a similarity matrix which is obtained by training and evaluating one-versus-all classifiers on validation set. So, this method quickly become impracticable because the cost of training OvA classifiers is too high for large-scale classification problem. In this paper, we introduce a new method to obtain a similarity matrix without using one-versus-all classifiers. To measure the similarity among classes, we used the sum-match kernel that is able to be calculated simply basing on the explicit feature map. Furthermore, to gain computational efficiency in classification, we also propose an algorithm for learning balanced label tree by balancing a number of class labels in each node. The experimental results on standard benchmark datasets ImageNet-1K, SUN-397 and Caltech-256 show that the performance of the proposed method outperforms significantly other methods.
Implementation of online meeting technology for socio - economic development in Vietnam
Tạp chí Kinh tế, Kinh doanh và Luật Đại học Quốc gia Thành phố Hồ Chí Minh - Tập 1 Số Q2 - Trang 70-78 - 2017
In recent years, Online Meeting Technology has been investigated intensively for possible use in various application areas e.g. Distance E-learning, Video Conferencing. The benefits of using Online Meeting Technology are to easily transfer the meeting information to peers around the world. For a company, using online meeting Technology can significantly reduce the company's costs associated with traveling expenses and lost working time. In this paper, we introduce a solution of setting online meeting room which is approaching the ordinary meeting room by using the online system “Easy Online” for helping companies to improve their management processes and training programs for their employees.
Applying OLAP technology to support decision making in sales process
Tạp chí Phát triển Khoa học và Công nghệ Đại học Quốc gia Thành phố Hồ Chí Minh - Tập 19 Số 2 - Trang 41-57 - 2016
OLAP (Online Analytical Processing) is a technology that enables the user to easily and selectively extract and view data from different points of view. It is also an important part of the decision system. This study proposes the decisions in fulfillment process which could be supported by the OLAP technology, including the quality of sales, the main product of the company, the salary and the bonus for sales staff, the credit limit and the price policy for the customer. Furthermore, this research also demonstrates the application of the OLAP technology in the decision making for the main product of A manufacturing and distributing company.
DLAFS CASCADE R-CNN: AN OBJECT DETECTOR BASED ON DYNAMIC LABEL ASSIGNMENT
Tạp chí tin học và điều khiển học - Tập 38 Số 2 - 2022
Object detection methods based on Deep Learning are the revolution of the Computer Vision field in general and object detection problems in particular. In detail, they are methods that belonged to the R - CNN family: Faster R - CNN and Cascade R - CNN. The characteristic of them is the Region Proposal Network, which is utilized for generating proposal regions that may include objects or not, then the proposals will be classified by the IoU threshold. In this study, we apply dynamic training, which adjusts this IoU threshold depending on the statistic of proposal regions on the Faster R - CNN and Cascade R - CNN, training on the SeaShips and DODV dataset. Cascade R - CNN with dynamic training achieve higher results compared to normal on both two datasets (higher 0.2% and 5.7% on the SeaShips and DODV dataset, respectively). In the DODV dataset, Faster R - CNN with dynamic training also perform higher results compared to its normal version, 4.4% higher.
#Object detection #Marine vehicle #Cascade R-CNN #Document detection.
Phrasal semantic distance for vietnamese textual document retrieval
Tạp chí tin học và điều khiển học - Tập 31 Số 3 - 2015
In this paper, a computational semantic method is proposed to estimate the phrasal semantic distance used in our model of a Vietnamese document retrieval system. The semantic distances between phrases are defined in terms of semantic classes and semantic relations to ensure that it can reflect how different two certain phrases are. To estimate the semantic distance, the semantic classes of a phase are identified by using the n-gram model. After identification of the semantic classes, their semantic relations are also identified by using a Vietnamese Lexicon Ontology. This handcrafted ontology contains defined semantic classes and their potential relations in Vietnamese language explicitly. For the evaluation purpose, a phrasal semantic retrieval system has been built to test with a data set of 720 phrases and 30 queries. The evaluation shows the precision of 96.6% and the recall of 78.4% on experimental results.
#Lexicon ontology #phrasal semantic analysis #semantic class #semantic distance #semantic information retrieval.
Tổng số: 44
- 1
- 2
- 3
- 4
- 5