Network traffic classification is crucial for cybersecurity, as it helps detect abnormal traffic and potential attack patterns, enabling timely protective measures. With the widespread adoption of remote work, video conferencing, and VPN services, the rapid growth of network traffic has increased the complexity of traffic classification. Traditional methods, such as port-based classification, Deep Packet Inspection, and Machine Learning (ML), face limitations in handling complex and encrypted traffic. In this study, we propose a novel network traffic classification system that combines 1D-CNN and Transformer, utilizing different activation functions to extract both local and long-range features. These complementary features are fused together to improve classification capability. The experimental results on the ISCX VPN-nonVPN dataset indicate that our proposed system outperforms baseline methods in most categories with respect to Precision, Recall, and F1-score.

Urbanization, characterized by the continuous growth of population and density in urban areas, has led to the expansion and increased complexity of transportation networks, exacerbating traffic congestion. Accurate traffic speed prediction is crucial for effective traffic network management and planning. As the complexity of real-world road networks increases, integrating spatial and temporal information for accurate traffic speed prediction has become a challenging research task. This paper proposes a novel approach by introducing a spatial-temporal graph neural network (STGNN)-based model to enhance the accuracy of traffic speed prediction. By employing an attention-based STGNN, we effectively capture the complex relationships among road segments in real-world scenarios. We utilize the Huber loss as the training objective to improve prediction accuracy. Furthermore, we present an architecture that incorporates Root Mean Square Layer Normalization into the Transformer and integrates the Spatial-Temporal Attention Wavenet (STAWnet) into the model backbone, enabling faster training while maintaining model stability. We evaluate the proposed model using five real-world traffic speed benchmark datasets. The experimental results demonstrate that our method achieves superior performance compared to state-of-the-art traffic speed prediction approaches.

In this study, we developed a tablet-based application, namely ubiquitous fraction (U-fraction) to assist elementary school students to learn fraction with authentic contextual support. Three topics of fraction included are fraction concept/comparison, simplification and addition/subtraction. An experiment was carried out to investigate the effects of students' learning behavior toward their fraction learning achievements using U-fraction, which allowed students to explore and verify fraction learning in authentic contexts. Totally, 80 fifth grade students were assigned and divided into three groups, one experimental group and two control groups. The first and second control group learned fraction using web-based application and traditional teaching method, while experimental group learned fraction using U-fraction. After the experiment, the results shown the experimental group performed better than those in both control groups. The reasons were that the graphical representation in authentic context facilitated students to learn fraction using U-fraction. Multimedia whiteboard like text and voice explanation provided by U-fraction also helped students to do more practices by themselves and fully explain their solutions. Furthermore, this study revealed that learning behavior variable i.e. note-taking (text), total fraction skill and assignment score variable had a significant correlation with post-test. We also found students' perception toward the system were acceptable.

This study aimed to investigate the relationship between learning behaviors (hypothesis-making, interpreting graphs, applying formula, conclusion-making and conceptual understanding) and learning achievement on vocational students who use our developed Ubiquitous-Physics (U-Physics) to learn simple pendulum in the experiment. Thus, this study conducted an experiment with participating second grade female vocational high school students who are less interests in physics, while hopefully using U-Physics in the physical experiment can motivate their interests and help their learning in physics. Yet, this study explored participants' perceptions towards the U-Physics. The results showed that there were significant correlation among hypothesis-making, interpreting graphs, applying formula, conclusion-making, conceptual understanding, and post-test. We also found that interpreting graphs and conceptual understanding were the two most important factors to affect learning achievement. In addition, students perceived that U-Physics was easy to use and useful for learning physics. Therefore, this study is strongly suggested experimental procedure should ask learners to interpret physics phenomena by graphs. Besides that, Teachers and researchers also should design physics learning activities to improve students' multiple representation skills supported by U-Physics system and utilize advanced features of mobile devices such as acceleration sensors, gyroscope, light sensors and GPS for learning physics.

In this paper, we examined users' behavior and perception in a mobile English learning application called "English Practice" (EP app). EP app has been placed on Google Play store over four years and has more than 800.000 installed. The study was conducted over two months from October 2015 to December 2015 through Google Analytic by drawing on random sample of 106,912 active users from 208 countries, especially 48,176 active users of ten countries included India, Philippine, Sudan, Turkey, US, Vietnam, Thailand, Pakistan, Malaysia and Egypt were analyzed. The results showed among top ten countries, India, Philippines and Malaysia seem the countries where users spent the longer duration per session as well as more screen views per session. Further understanding about English language status among top ten countries, English popularity inside domestic was likely one of the reasons to explain the difference. In addition, the similar trend found among ten countries in which Quiz function seems been visited mostly, followed by Chat-room and Flashcard. Grammar lessons were the less popular function. We suggested that simplify learning task and reducing the content presenting in each session might look more apparel to users of EP app.

In this study, we evaluated the effect of push notifications on mobile app engagement. Over the duration of a month, we collected data from 12,071 online users of an English learning app named 'English Practice' which can be freely installed from the Google Play store. Four criteria of app engagement used by us were as follows: session count, session length, time consumption and retention. In addition, we examined the frequency of sending notifications, how to display notifications and the use of notifications on wearable devices. The results show that push notifications with regards to incoming messages can increase learner engagement most. On the other hand, abusing push notifications (e.g. if they are too frequent) result in the opposite effect. Based on the results, we suggest that educators or developers focus on an increased integration of social communication tools inside apps.

This study aimed to explore the effectiveness of applying the speech-to-text recognition (STR) technology during lectures in English on learning performance of non-native English students. Furthermore, how useful texts generated by the STR for students of different English as foreign language (EFL) ability during different difficulty level lectures was explored. For this reason, one experiment was carried out in this study. Two lectures, one is intermediate difficulty level lecture and the other is advanced, were administered in a traditional classroom and the STR was adopted to aid students' learning. Results of this study showed that students who used STR-generated texts outperformed students who did not use them. According to students, STR-texts were useful to follow the instructor, clarify some words, confirm what the lecturer has said, make up missed information, and enhance understanding. It was found that STR-texts were beneficial for learning of low EFL ability students during both lectures while high EFL ability students only used STR-texts during the advanced level lecture. Based on these results, several suggestions and implications for teaching and research community are proposed in this study.

This study aimed to evaluate how differently co-located users perform and interact while solving problems by using existing and newly developed interaction mechanisms of a multi-touch tabletop system on a large-size tabletop display. Furthermore, this study aimed to investigate users’ perceived situation awareness due to the system. This study conducted one experiment and introduced three newly developed interaction mechanisms: two-way rubber band, drag-and-response, and centre collection area. Two-way rubber band enables users not only to send but also to request digital objects. Drag-and-response allows users to accept or to reject digital objects sent by others; therefore, it eliminates conflicting situations and interference. Centre collection area stores, organises, and manages shared digital resources on a surface of a tabletop display, so that users have an overall picture of available resources. This paper discusses results of an experiment, research findings, and implications along with conclusions and several suggestions for future development and research.

One major drawback of previous research on speech-to-text recognition (STR) is that most of their findings showing an effectiveness of STR for learning in traditional classroom were based on subjective evidences. Very few studies have used eye-tracking techniques to investigate visual attention of students on STR generated text. Furthermore, not much attention was paid to learning differences (i.e. Learning ability) to use STR-text. Therefore, this study carried out one experiment in which participants' visual attention on STR generated text during lectures was investigated by employing eye-tracking technique. Besides, how differently effective STR-text can be to influence participants' learning achievement was tested. Furthermore, this paper discusses results, research findings, and implications along with conclusions and several suggestions for future development and research.

This study aimed to investigate effects of unidirectional versus reciprocal teaching strategies on computer programming learning. Thus, this study conducted an experiment with participating novice programmers, who applied no strategies, unidirectional, and reciprocal teaching strategies for computer programming learning by using a web-based learning system (VPen). Yet, this study explored participants' perceptions toward the VPen system and the strategies. The paper reports results of this study as well as it provides useful suggestions on application of the strategies by using the VPen for teaching and research community of this field and for future studies.

Prior studies show that active-motion early after a fracture surgery can effectively reduce the tissue viscosity. However, due to the postoperative pain and unawareness of the postoperative limitations, a patient prefers to take medication to alleviate pain while complicated rehabilitation exercises can only be implemented under the limited intervention of a therapist, resulting in the patient's lack of interactive feedback on a daily basis and a failure of self-motions to miss the timing of rehabilitation. This study is intended to develop a lower limb fracture postoperative-guided interactive rehabilitation training system for the hip, knee and ankle joints, and establish a method of motion analysis and a method of motion performance assessment in conjunction with the wireless sensor technology and animation techniques, providing an objective assessment from the caregiver and learning the patient's rehabilitation status with a simple and easy recording means. Additionally, with the design of interactive feedback, the improved postoperative rehabilitations are tried to increase the participant's interest in self-management of health and facilitate to keep tracking the rehabilitation information. The medical evidences show that a guided rehabilitation system can enhance the patient's motive and willingness to be committed to a rehabilitation training as well as increase Quality and Amount of the exercise activity during the training process.

In this study we proposed self-assessment with prompts, learning journal and referencing through sharing learning methods for regulation of cognition with one annotation system called VPen. We also aimed to investigate the effectiveness of the learning methods on web-based programming learning performance and achievement. The results of this study revealed students perceived the learning methods are useful for the strategies of regulation of cognition. We also found that self-assessment with prompts, learning journal and referencing through sharing of learning performance and behavior significantly correlate with each other and learning achievement. In addition, the results revealed learning journal brings more effect on learning performance and achievement than self-assessment with prompts.

Synchronous discussion and asynchronous discussion have different advantages and disadvantages. However, it is unclear how students react to these two models. Thus, we developed a learning environment that combines these two models, and then examined how students selected them. The findings show that students selected different models at different discussion stages. Additionally, the mixed discussion model, which combines synchronous and asynchronous models, can achieve better performance in discussion, regardless of the types of discussion activity.

Wireless Sensor Network is one of the 21 ideals in 21 century, and health care is one of its many applications. Generally speaking, there are two keys for the successful construction of Wireless Sensor Network, one it the editing of appropriate data transfer protocol at the front end of the network, the next is the possession of robust database and effective network platform in the back end that can meet the need of the users. Based on the above two requirements, the customized wireless physiological sensor network designed by this research is based on the real environment of health care room, under the premise of "customer is number 1", comfortable equipment, effective database and simple network system are used to achieve the objective of omnipresent health care. Based on the experience and technology obtained from this real application, we hope to build more complicated U-healthcare system to be used by the health service center.

This study aimed to apply Speech to Text Recognition (STR) for individual oral presentations and group discussions of students in a synchronous cyber classroom. An experiment was conducted to analyze the effectiveness of applying STR on learning performance. Students' perceptions and behavioral intentions toward using STR were also investigated. The results revealed students of the experimental group performed significantly better compared to students of the control group in writing essays, intermediate test and post-test. In addition, students perceived the STR was easy to use and useful for learning. Most of the students expressed that they are willing to use the STR for learning in the future. The results of this study showed that the STR is beneficial to students' oral presentations and group discussions in a synchronous cyber classroom so as to improve their overall learning achievement.

To improve students' interaction in a collaborative learning environment to complete their tasks is very critical for developing students' social interaction. This study designed a system of Multi-robots to discuss processes of students' collaborative strategies and behavioral interaction by controlling robots under different scenarios to complete the dice-pushing tasks. As a result, students developed three collaborative strategies: independent-control, mutual-control and coordinator-directed. Students with independent-control strategy spent most time completing the task and those with mutual-control strategy spent least time. Additionally students had more behavioral interaction in mutual-controlled strategy. Regarding to learn how to learn “collaborative learning”, mutual-control strategy was formed under the scenario without evenly distributed resources. Therefore it is suggested that in order to increase students' interaction, the related collaborative activities could be designed based on this aspect.

In the study, one Web annotation system VPen was proposed for facilitating students to foster metacognition by reviewing self and otherspsila annotation and homework. Annotations are personally meaningful and useful for individual reflection later, while homework sharing is generally useful for all because it is assigned by instructors. Therefore, through reviewing these two products, annotation and homework, some metacognition may occur. We conducted the quasi experiment employing VPen system to study the influence of reviewing annotation and homework on math learning. The results showed that reviewing selfpsilas annotation and homework had significant influence on learning achievements. Regarding consulting otherspsila annotation and homework, some interesting different phenomena were obtained; for consulting otherspsila annotation, less usefulness was found on learning achievement because annotation is usually meaningful and useful to it owner only, while consulting otherspsila homework is more useful because all students need to solve homework and can get much benefits from studying other solving methods. From this study, we suggested that the design of learning activities should consider not only taking annotation and homework but also promoting students to review them in order to foster metacognition.

The purpose of this study is to deepen our understanding of communication by proposing a related, novel construct: intra-action. We define intra-action as communication with oneself through the use of external tools. In intra-action, similarly to communication involving distinct parties, learners use discourse conducted with support from communication technology to deepen their understanding of the subject matter. Unlike in conventional communication in which the discourse is targeted at others, in intra-action the discourse is targeted at the learners themselves. Examples of intra-action activities and tools are digital annotation, electronic diaries and blogs. We study its impact on learning outcomes and its relationships with other kinds of communication: interaction (substantive communication with others) and outeraction (rapport-building communication with others). Three tools: annotation, discussion board and instant messenger were designed to support intra-action, interaction and outeraction, respectively. After our experiment, the results clearly demonstrate that intra-action has stronger relationship with learning outcomes than interaction or outeraction. Also, considerable correlation was detected between intra-action and interaction. Subsequent interviews with learners indicated that considerable part of the positive relationship between intra-action and learning achievement may be due to cause-effect, namely, intra-action activities are perceived by learners as contributing to positive learning outcomes.

Web-based programming has become a popular and vital issue in recent years. The rapid growth of various applications not only demonstrates the importance of web-based programming, but also reveals the difficulty of training relevant skills. The difficulty is owing to the lack of facilities such as online coding, debugging and peer help to assist the students in promoting their cognitive development in web-based programming. To cope with these problems, in this paper, a web-based programming assisted system, “WPAS”, is proposed, which is able to support five programming activities with various difficulty levels of cognition based on Bloom’s cognitive taxonomy. WPAS provides online coding, debugging and annotation tools to conduct the training and peer assessment for web-based programming. Experimental results of 47 undergraduate students show that the innovative approach is helpful to the students in improving their cognitive development in Web-based programming. In addition, according to the results of the questionnaire, most of the participants perceived the ease of use and usefulness of the proposed system. Therefore, this study suggests that teachers could design Web-based programming activities supported by the WPAS system to improve students’ cognitive development in web-based programming.

The purpose of this paper is to promote a new online mathematical learning model that students not only use electronic blackboard to write down their mathematical problem solving solutions but also use the voice recording tool to give oral explanations about their thinking behind the solutions. To cultivate students' critical thinking capability and encourage collaborative peer learning, the new learning model also requests students to criticize others' solutions and to reply others' arguments. With the multimedia supporting tools, students are more easily to communicate with each other about what they think and how they solve mathematical problems by using the electronic blackboard and voice recording tools. After analyzed the recorded information with statistical method, we found that the performance of female students is superior to those male students in communications and mathematical problem solving. Additionally, students with higher final exam grades have better mathematical abilities of doing critiques, arguments and communications.

We propose an efficient multicast cell-scheduling algorithm, called multiple-slot cell-scheduling algorithm, for multicast ATM switching systems with input queues. Cells in an input-queueing system are usually served based on the first-in-first-out (FIFO) discipline, which may have a serious head-of-line (HOL) blocking problem. Our algorithm differs from previous algorithms in that we consider the output contention resolution for multiple time slots instead of the current time slot only. Like a window-based scheduling algorithm, our algorithm allows cells behind an HOL cell to be transmitted prior to the HOL cell in the same input port. Thus, HOL blocking can be alleviated. We have illustrated that the delay-throughput performance of our algorithm outperforms most of those algorithms that consider only the output contention resolution for the current time slot. We also present a simple and efficient architecture for realizing our algorithm, which can dramatically reduce the time complexity. We believe that the proposed architecture is very suitable for multicast asynchronous transfer mode (ATM) switching systems with input queues.

Sharing buffer space between switch ports greatly improves the performance of the switching systems. However sharing buffers in a fair manner is not an easy task. In this paper we propose a high-performance ATM switching system with fairly and completely shared buffers. The core of the proposed switching system is a novel buffer management mechanism with which buffer space in the switching system can be completely shared by all the switch ports in a fair manner. The proposed buffer management mechanism works based on a simple algorithm and can be easily implemented by hardware. The performance of the proposed switching system is evaluated by both analytical model and simulation. The results show that high buffer utilization and low packet lost ratio can be achieved under both uniform and nonuniform traffic loads.