Although online learning platforms are gradually becoming commonplace in modern society, learners’ high dropout rates and serious academic performance require more attention within the virtual learning environment (VLE). This study aims to predict students’ performance in a specific course as it is continuously running, using the statistic personal biographical information and sequential behavior data with VLE. To achieve this goal, a novel recurrent neural network (RNN)-gated recurrent unit (GRU) joint neural network is proposed to fit both static and sequential data, where the data completion mechanism is also adopted to fill the missing stream data. To incorporate the sequential relationship of learning data, three kinds of time-series deep neural network algorithms: simple RNN, GRU, and LSTM are first taken into consideration as baseline models. Their performances are compared in identifying at-risk students. Experimental results on Open University Learning Analytics Dataset (OULAD) show that simple methods like GRU and simple RNN have better results than the relatively complex LSTM model. The results also reveal that different models have different peak performance time, which results in the proposed joint model that achieves over 80% prediction accuracy of at-risk students at the end of the semester.

Personalized learning paths aim to save learning time and improve learning achievements by providing the most appropriate learning sequence for heterogeneous students. Most existing methods that construct personalized learning paths focus on students’ characteristics or knowledge structure, while ignoring the critical roles of learning states. This study describes a dynamic personalized learning path planning algorithm to recommend appropriate knowledge points for online students based on their learning states and the difficulty of each knowledge point. The proposed method first calculates the difficulty of knowledge points automatically and constructs a knowledge difficulty model. Then, a dynamic knowledge mastery model is built based on learning behavior and normalized test scores. Finally, a path that satisfies students’ personalized changing states is generated. To achieve the aforementioned goal, a novel method that calculates the difficulty of knowledge points automatically is proposed. Moreover, the personalized learning path planning method proposed in this research is not limited to a particular course. To evaluate the method, we use a series of approaches to verify the impact of the personalized path on student learning. The experimental results demonstrate that the proposed algorithm can effectively generate personalized learning paths. Results demonstrate that the personalized path proposed by the algorithm can improve effective behavior rates, course completion rates and learning efficiency. Results also show that the personalized learning paths based on student states would help students to master knowledge.

COVID-19 has affected traditional instructional activities. Home-based isolation and restrictive movement measures have forced most learning activities to move from an offline to an online environment. Multiple studies have also demonstrated that teaching with virtual tools during the COVID-19 pandemic is always ineffective. This study examines the different characteristics and challenges that virtual tools brought to online education in the pre-pandemic and pandemic era, with the aim of providing experience of how virtual tools supported purely online learning during a health crisis. By searching keywords in public databases and review publications, this study tries to summarize the major topics related to the research theme. These topics are the characteristics of learning supported by technologies in pre-pandemic and pandemic era, the challenges that education systems have faced during the COVID-19 pandemic. This study also compares the functions, advantages and limitations of typical virtual tools, which has rarely been done in previous studies. This study tries to present the features of virtual tools that support online learning and the challenges regarding real-life risk scenarios, and tries to provide educational institutions with a distinct perspective for efficient teaching and learning in future potential health crises.

Recently, deep learning has enabled a huge leap forward in image inpainting. However, due to the memory and computational limitation, most existing methods are able to handle only low-resolution inputs, typically less than 1 K. With the improvement of Internet transmission capacity and mobile device cameras, the resolution of image and video sources available to users via the cloud or locally is increasing. For high-resolution images, the common inpainting methods simply upsample the inpainted result of the shrinked image to yield a blurry result. In recent years, there is an urgent need to reconstruct the missing high-frequency information in high-resolution images and generate sharp texture details. Hence, we propose a general deep learning framework for high-resolution image inpainting, which first hallucinates a semantically continuous blurred result using low-resolution inpainting and suppresses computational overhead. Then the sharp high-frequency details with original resolution are reconstructed using super-resolution refinement. Experimentally, our method achieves inspiring inpainting quality on 2K and 4K resolution images, ahead of the state-of-the-art high-resolution inpainting technique. This framework is expected to be popularized for high-resolution image editing tasks on personal computers and mobile devices in the future.

Subtitles are crucial for video content understanding. However, a large amount of videos have only burned-in, hardcoded subtitles that prevent video re-editing, translation, etc. In this paper, we construct a deep-learning-based system for the inverse conversion of a burned-in subtitle video to a subtitle file and an inpainted video, by coupling three deep neural networks (CTPN, CRNN, and EdgeConnect). We evaluated the performance of the proposed method and found that the deep learning method achieved high-precision separation of the subtitles and video frames and significantly improved the video inpainting results compared to the existing methods. This research fills a gap in the application of deep learning to burned-in subtitle video reconstruction and is expected to be widely applied in the reconstruction and re-editing of videos with subtitles, advertisements, logos, and other occlusions.