Search

In Vietnam’s society where the economy is under restructuring into a market one, the renovation of education started in 1982 and the comprehensive educational renovation have created huge changes in the national educational system in 2001. The educational renovation has been led by the Vietnam Communist Party’s (CPV) policy course, regulated by the country’s legislation and has opened up education opportunities, reduced social inequality in education, and changing educational level structure. A number of studies have clarified factors of educational opportunities and their relationship to the human resource and economy, but these studies are relevant mainly to societies with developed and developing market economy. Therefore, based on the systematical theory approach and literature analysis method, this paper concentrates on clarifying the education renovation triangle comprising education policy, education opportunity, and educational level in a society restructuring into a market economy in Vietnam.

Opinion mining (also known as sentiment analysis) through customers’ reviews or feedbacks which can identify the user opinion about different product features has received a lot of attention exhibited in numerous studies. The majority of recommender systems have recommended products based only on overall evaluation and primarily based on expert’s evaluation. In this work, we propose a method to explore Vietnamese reviews extracted from e-commerce websites in Vietnam to provide suggestions in products selection based on products’ features/functions. The proposed approach introduces a topic-based model to identify products’ features which are mentioned in customer comments/reviews. The proposed system is implemented with the integration combining the VietSentiWordnet to calculate the importance scores for the features of each product. We also construct a product recommendation database which can store customers preference and purchases history. The work is analysed on more than 2,000 Vietnamese comments/reviews about laptop products and is expected to be feasible to apply in practical cases.

Recent advances in deep learning facilitate the development of end-to-end Vietnamese text-to-speech (TTS) systems that produce Vietnamese voices with high intelligibility and naturalness. However, enabling these systems to speak Vietnamese and English words in the same utterance fluently remains a challenge known as the code-switching (CS) problem in speech synthesis. The main reason is that it is not easy to obtain a large amount of high-quality CS corpus from a Vietnamese speaker. In this paper, we explore the efficacy of three approaches, which are based on the Tacotron-2 end-to-end framework, to build such a Vietnamese TTS system under a limited code-switched data scenario: (1) CS synthesis based on grapheme-to-syllable (G2S), (2) CS synthesis based on speaker embedding, and (3) CS synthesis based on speaker embedding and language embedding. We handle English and Vietnamese words in the code-switched input text by converting them into Vietnamese syllables using our G2S model. For the speaker-embedding based approach, we combine Vietnamese monolingual data in our dataset with an English public dataset to train a multi-speaker Tacotron-2 system. The experimental results show that adding language embedding is effective, and training with character input representations outperforms phonemes. Thus, the speaker and language-embedding based approach achieves strong results in naturalness for CS speech. Besides, the G2S-based CS synthesis also has good results, with almost absolute English pronunciation accuracy.

Text normalization for Text-To-Speech includes several challenging tasks in natural language processing such as verbalizing abbreviations, Out-Of-Vocabulary (OOV) words, and chunking phrases for long sentences without punctuation. Instead of dealing with these tasks independently, we propose a multi-task end-to-end model based on a denoising auto-encoder. As enriching information via multi-task learning and text denoising, the proposed approach shows improvement in the text normalization task, especially for complicated cases which require context and language understanding. In addition, we also design a novel process of data prepossessing to leverage annotated data for training. According to experiments on a handcrafted test set of 200,000 sentences in Vietnamese, our model achieves an overall accuracy of 95.5%.

The priority-based traffic light controlling system at an intersection dedicates right-of-way to the highest priority direction. The system is applied for urban traffic system in Vietnam, where various types of vehicles run in close proximity to one another, needs a traffic model suitable to determine the priority of directions going through intersection. The paper has studied current traffic models, microscopic and macroscopic traffic flow models as well as three-phase traffic model, to propose spatio-temporal traffic statistics model. The paper considers red light as a traffic obstacle causing moving jam behind traffic light. The proposed spatio-temporal traffic statistics model studies traffic flow on a road segment bounded by two intersections, where various types of vehicles travel. The spatio-temporal traffic statistics model represents the road surface area occupied by vehicles at locations and over time as spatio-temporal histograms. The expectation of histogram is a factor determining the priority of direction going through intersection.

This paper utilizes the traffic flow model of car-following to study the movement of two consecutive vehicles travelling in the same lane of a freeway. The car-following models defined the speed and headway of a moving vehicle but have not analyzed their relation. In the proposed block-moving approach, the following vehicle is considered as a moving object obstructed by another one moving in the same direction. In that, the correlation between the speed and headway of the following vehicle, which are recorded discretely in real time by instruments installed on the following vehicle, is analyzed and processed to avoid collisions caused by the following vehicle. Additionally, the following vehicle must react accordingly to the behaviors of the leading vehicle by increasing or decreasing speed to a desired speed. This study proposed a methodology to effectively determine how much engine or brake force should be made to achieve the desired speed.

Current bus maps are created by tracing ground trajectories of bus routes on available base maps. These bus maps provide with the spatial information of ground trajectories and bus stops; some websites have associated departure time table with each ground trajectory on the map. It is considered that bus users need not only spatial but also temporal information. This paper integrates visualization with temporal geography to represent spatio-temporal data of bus network as a visualized-space-time bus map. The visualized-space-time bus map displays visually the movement data of bus as graphs of bus routes and bus trips on screen. The graphs are classified as visual classes according to their attributes. The visual features of the graphs are enhanced by integrating graph classes with retinal variables. The visualized-space-time bus map enables users to analyze visually the spatio-temporal data of bus network for their travels.

In Hanoi, Vietnam, the people heavily depend on groundwater resources for both domestic and industrial uses due to the seriously polluted problems occurred in most of the river streams system. Considering the social dimension in the context of sustainable development is a key issue in groundwater resources management. The main object of this study is to assess the social sustainability of groundwater resources in Hanoi by applying a simple Analytical Hierarchy Process (AHP) approach. By applying the simple AHP, the most important task is to comprehensively study the current social situation and actual problems to define the appropriate foremost components contributing to the sustainability goal. In this study, we successfully not only select the appropriate list of three main social sustainability aspects including the quantity, quality and management aspects and the twelve core social sustainability indicators but also appropriately assessing the sustainability of the groundwater resources from the social point of view. By gathering the data from trusted sources, the groundwater quantity, quality, and management aspects are socially assessed at a good level of 0.65, a good level of 0.61 and a slightly acceptable level of 0.46, respectively. The composite social sustainability index, S, is assessed at a strongly acceptable level of 0.57, indicating that Hanoi’s groundwater resource is socially acceptable. The results of SSA are not only sensitive to data availability but also the definition of the indicators. The linear relationships do not reflect the actual situation well and the non-linear ones could be used to replace these sustainability scales of the actual values. The social sustainability indices are evaluated at the moderately high values, improbably reflecting the current problems in the target area. The study then points out the main research gaps and suggests the ways to improving social sustainability assessment for the better groundwater resource development.