Bio: Daqian Shi is an academic researcher from University of Trento. The author has contributed to research in topics: Computer science & Character (mathematics). The author has an hindex of 3, co-authored 6 publications receiving 43 citations. Previous affiliations of Daqian Shi include Jilin University & University of Manchester.
TL;DR: A learning path recommendation model is designed for satisfying different learning needs based on the multidimensional knowledge graph framework, which can generate and recommend customized learning paths according to the e-learner’s target learning object.
Abstract: E-learners face a large amount of fragmented learning content during e-learning. How to extract and organize this learning content is the key to achieving the established learning target, especially for non-experts. Reasonably arranging the order of the learning objects to generate a well-defined learning path can help the e-learner complete the learning target efficiently and systematically. Currently, knowledge-graph-based learning path recommendation algorithms are attracting the attention of researchers in this field. However, these methods only connect learning objects using single relationships, which cannot generate diverse learning paths to satisfy different learning needs in practice. To overcome this challenge, this paper proposes a learning path recommendation model based on a multidimensional knowledge graph framework. The main contributions of this paper are as follows. Firstly, we have designed a multidimensional knowledge graph framework that separately stores learning objects organized in several classes. Then, we have proposed six main semantic relationships between learning objects in the knowledge graph. Secondly, a learning path recommendation model is designed for satisfying different learning needs based on the multidimensional knowledge graph framework, which can generate and recommend customized learning paths according to the e-learner’s target learning object. The experiment results indicate that the proposed model can generate and recommend qualified personalized learning paths to improve the learning experiences of e-learners.
TL;DR: The knowledge representation learning approach is used in the MRP2Rec method to learn and represent multiple-step relation paths, and they are further utilized to generate prediction lists by inner products in top- $K$ recommendations.
Abstract: Knowledge graphs (KGs) have been proven to be effective for improving the performance of recommender systems. KGs can store rich side information and relieve the data sparsity problem. There are many linked attributes between entity pairs (e.g., items and users) in KGs, which can be called multiple-step relation paths. Existing methods do not sufficiently exploit the information encoded in KGs. In this paper, we propose MRP2Rec to explore various semantic relations in multiple-step relation paths to improve recommendation performance. The knowledge representation learning approach is used in our method to learn and represent multiple-step relation paths, and they are further utilized to generate prediction lists by inner products in top-K recommendations. Experiments on two real-world datasets demonstrate that our model achieves higher performance compared with many state-of-the-art baselines.
TL;DR: A deep learning-based visual emotion recognition system to recognize emotions of all students in the classroom and reduces response time to achieve real-time recognition and intervention in negative emotional contagion in classroom.
Abstract: Literature has indicated that negative emotions may lead students to disengagement in teaching activities. Furthermore, the contagion of negative emotion is similar to infectious disease diffusion that drives more students into negative emotions. However, few methods have been brought forward to intervene in negative emotional contagion in real time, and most of them are limited to interventions of teachers, which are often not timely and even cause students to resist. Intervention in negative emotional contagion in classroom imposes several fundamental challenges on model and system design. In this paper, we address these issues from the following three aspects: (1) to design an emotional contagion model for classroom scene to locate the source of negative emotional contagion; (2) to develop deep learning-based visual emotion recognition system to recognize emotions of all students in the classroom; (3) to design and deploy the emotion recognition system as an edge computing-based service for minimizing response time to achieve multi-person emotional recognition and intervene in real time. We have applied the system to real-world classroom. Our results have shown that the system achieved two objectives: (1) reducing the number of students with negative emotions; (2) reducing response time to achieve real-time recognition and intervention. Meanwhile, this work provides a new perspective on research into emotional contagion in the classroom and smart education.
TL;DR: The results show that the gamification method of “splitting and combining” (SC) works well in promoting learning enjoyment and that demonstrates better performance than combining.
Abstract: The understanding of the structure of knowledge is an essential step of education. Although teachers offer the information foundation and relationship among knowledge points, there are still few methods to encourage students to explore the structure of knowledge by themselves outside of classes. This paper explores the gamification method and the knowledge structure of computer science. We assess the gamification method of “splitting and combining” (SC) to encourage students to finish the process of learning structured knowledge in the university. The results show that this method works well in promoting learning enjoyment and that splitting demonstrates better performance than combining. We can consider the SC method when recommending a gamification method to engage students in structural learning assistance in future smart university education.
TL;DR: A zero-shot character recognition framework via radical extraction, i.e., REZCR, is proposed to improve the recognition performance of few-sample character categories, in which information on radicals is exploited by de- composing and reconstructing characters following orthog-raphy.
Abstract: The long-tail effect is a common issue that limits the perfor- mance of deep learning models on real-world datasets. Character image dataset development is also affected by such un- balanced data distribution due to differences in character usage frequency. Thus, current character recognition methods are limited when applying to real-world datasets, in partic-ular to the character categories in the tail which are lack- ing training samples, e.g., uncommon characters, or characters from historical documents. In this paper, we propose a zero-shot character recognition framework via radical extraction, i.e., REZCR, to improve the recognition performance of few-sample character categories, in which we exploit information on radicals, the graphical units of characters, by de- composing and reconstructing characters following orthog-raphy. REZCR consists of an attention-based radical infor- mation extractor (RIE) and a knowledge graph-based character reasoner (KGR). The RIE aims to recognize candidate radicals and their possible structural relations from character images. The results will be fed into KGR to recognize the target character by reasoning with a pre-designed character knowledge graph. We validate our method on multiple datasets, REZCR shows promising experimental results, especially for few-sample character datasets.
TL;DR: The TAM-based proposed scale has been successfully explained factors predicting the use of e-learning among Indonesian sport science students during the pandemic and the finding of significant relationships between facilitating condition and perceived ease of use and between facilitatingcondition and perceived usefulness was reported.
Abstract: This study was to explore factors predicting the use of e-learning during Corona Virus Disease 2019 (Covid-19) among sport science education students In Indonesia Higher Education Institutions (HEIs). The study was conducted through survey with 974 participating students from five Indonesian HEIs. An extended Technology Acceptance Model (TAM) with facilitating condition as the external factor was implemented to be the theoretical framework of this study. An analysis method through Partial Least Squares Structural Equation Modeling (PLS-SEM) was employed to measure and assess the proposed model. The findings informed that: (1) the TAM-based proposed scale has been successfully explained factors predicting the use of e-learning among Indonesian sport science students during the pandemic; (2) the finding of significant relationships between facilitating condition and perceived ease of use and between facilitating condition and perceived usefulness was reported; and (3) the significant relationships among core components of TAM were found except for one, relationship between perceived usefulness and attitude.
TL;DR: This survey is the first to provide an inclusive definition to the notion of domain KG, and a comprehensive review of the state-of-the-art approaches drawn from academic works relevant to seven dissimilar domains of knowledge is provided.
Abstract: Knowledge Graphs (KGs) have made a qualitative leap and effected a real revolution in knowledge representation. This is leveraged by the underlying structure of the KG which underpins a better comprehension, reasoning and interpretation of knowledge for both human and machine. Therefore, KGs continue to be used as the main means of tackling a plethora of real-life problems in various domains. However, there is no consensus in regard to a plausible and inclusive definition of a domain-specific KG. Further, in conjunction with several limitations and deficiencies, various domain-specific KG construction approaches are far from perfect. This survey is the first to offer a comprehensive definition of a domain-specific KG. Also, the paper presents a thorough review of the state-of-the-art approaches drawn from academic works relevant to seven domains of knowledge. An examination of current approaches reveals a range of limitations and deficiencies. At the same time, uncharted territories on the research map are highlighted to tackle extant issues in the literature and point to directions for future research.
TL;DR: The most significant challenges of the methods that are applied to personalize learning paths need to be tackled in order to enhance the quality of the personalization.
Abstract: A learning path is the implementation of a curriculum design. It consists of a set of learning activities that help users achieve particular learning goals. Personalizing these paths became a significant task due to differences in users’ limitations, backgrounds, goals, etc. Since the last decade, researchers have proposed a variety of learning path personalization methods using different techniques and approaches. In this paper, we present an overview of the methods that are applied to personalize learning paths as well as their advantages and disadvantages. The main parameters for personalizing learning paths are also described. In addition, we present approaches that are used to evaluate path personalization methods. Finally, we highlight the most significant challenges of these methods, which need to be tackled in order to enhance the quality of the personalization.
TL;DR: The design of a scientific publication management model to integrate scientific metadata based on the knowledge graph and data analysis technologies is presented to enhance scientific retrieval efficiency and reduce learning difficulty in the scientific domains and encourage non-researchers to utilize scientific resources in their study and work.
Abstract: In recent years, with the rapid growth of science and innovation, plenty of constantly-updated scientific achievements containing innovative knowledge can be acquired and used to solve problems. However, most undergraduate students and non-researchers cannot use them efficiently. In traditional teacher-centric education, education for sustainability is often marginalized and the interdisciplinary demand is neglected. Additionally, it fails to provide education for learners to connect abstract knowledge with actual world problems. This paper presents the design of a scientific publication management model to integrate scientific metadata based on the knowledge graph and data analysis technologies. Based on this model, an interdisciplinary transregional multiple application platform could be realized for scientific resource retrieval and analysis, the purpose of which is to enhance scientific retrieval efficiency and reduce learning difficulty in the scientific domains and encourage non-researchers to utilize scientific resources in their study and work. Finally, to evaluate this model, the use of the case of an entrepreneurship scientific publication management prototype system was implemented. This work not only favors student’s learning for sustainability through analysis and knowledge management functions, but also promotes their awareness, comprehensive thinking, and the skills to deal with the issues of sustainability in their future work.
TL;DR: A novel model named Multi-Channel Attention Selection Generative Adversarial Network (SelectionGAN) for guided image-to-image translation, where an input image is translated into another while respecting an external semantic guidance.
Abstract: We propose a novel model named Multi-Channel Attention Selection Generative Adversarial Network (SelectionGAN) for guided image-to-image translation, where we translate an input image into another while respecting an external semantic guidance. The proposed SelectionGAN explicitly utilizes the semantic guidance information and consists of two stages. In the first stage, the input image and the conditional semantic guidance are fed into a cycled semantic-guided generation network to produce initial coarse results. In the second stage, we refine the initial results by using the proposed multi-scale spatial pooling \& channel selection module and the multi-channel attention selection module. Moreover, uncertainty maps automatically learned from attention maps are used to guide the pixel loss for better network optimization. Exhaustive experiments on four challenging guided image-to-image translation tasks (face, hand, body and street view) demonstrate that our SelectionGAN is able to generate significantly better results than the state-of-the-art methods. Meanwhile, the proposed framework and modules are unified solutions and can be applied to solve other generation tasks, such as semantic image synthesis. The code is available at this https URL.