During the COVID-19 epidemic, most programming courses were revised to distance learning. However, many problems occurred, such as students pretending to be actively learning while actually being absent and students engaging in plagiarism. In most existing systems, obtaining status updates on the progress of a student’s learning is hard. In this paper, we first define the term “class loyalty”, which means that a student studies hard and is willing to learn without using any tricks. Then, we propose a novel method combined with the parsing trees of program codes and the fuzzy membership function to detect plagiarism. Additionally, the fuzzy membership functions combined with a convolution neural network (CNN) are used to predict which students obtain high scores and high class loyalty. Two hundred and twenty-six students were involved in the experiments. The dataset was randomly separated into the training datasets and the test datasets for twenty runs. The average accuracies of the experiment in predicting which students obtain high scores using the fuzzy membership function combined with a CNN and using the duration and number of actions are 93.34% and 92.62%. The average accuracies of the experiment in predicting which students have high class loyalty are 95.00% and 92.74%. Both experiments show that our proposed method not only can detect plagiarism but also can be used to detect which students are diligent.

The fermentation of Lactiplantibacillus plantarum is a complex bioprocess due to the nonlinear and dynamic nature of microbial growth. Traditional monitoring methods often fail to provide early and actionable insights into fermentation outcomes. This study proposes a deep learning-based predictive system using convolutional neural networks (CNNs) to classify fermentation trajectories and anticipate final cell counts based on the first 24 h of process data. A total of 52 fermentation runs were conducted, during which real-time parameters, including pH, temperature, and dissolved oxygen, were continuously recorded and transformed into time-series feature vectors. After rigorous preprocessing and feature selection, the CNN was trained to classify fermentation outcomes into three categories: successful, semi-successful, and failed batches. The model achieved a classification accuracy of 97.87%, outperforming benchmark models such as LSTM and XGBoost. Validation experiments demonstrated the model’s practical utility: early predictions enabled timely manual interventions that effectively prevented batch failures or improved suboptimal fermentations. These findings suggest that deep learning provides a robust and scalable framework for real-time fermentation control, with significant implications for enhancing efficiency and reducing costs in industrial probiotic production.

In recent years, bilingual education has become a critical index, and with its growth, internationalization has been significantly improved in the Republic of China (Taiwan). Therefore, countries worldwide are promoting all-English teaching, taking English as a medium of instruction and indicator of university education while moving toward an internationalized curriculum and teaching excellence. Finding the required number of bilingual teachers is also one of the keys to increasing the index. If information technology can be adopted to reduce the time and cost of bilingual teacher training, it will significantly improve the effectiveness of bilingual education promotion. However, the current traditional online training system only focuses on the delivery of one-way training content, and it cannot assist in judging the training effectiveness of the trainees. In this paper, we integrate the pose estimation technology into the online teacher training system to analyze the interactive content of the trainees during the training. We can assist in recording and interpreting the teaching demonstration process in classroom observation. We also verify the result by comparing our method’s efficiency and judgment accuracy with the traditional way. The results show that our approach is more convenient and cost-efficient.

This study explores the challenges faced by students in higher education and their learning conditions. Today, higher education has become a crucial element in our social development. This study utilizes the DICE system for simulating, recording, predicting, and analyzing students' performance in the General English Proficiency Test (GEPT). This study initially collects data from the server, subsequently utilizing XGBOOST to forecast students' score categories. Furthermore, an exploration of students' learning patterns is conducted, considering time and score distributions. We amassed simulation scores from 1631 students in the GEPT. The outcomes indicate an 85% accuracy in predicting students' score categories using our system. Additionally, we analyzed the temporal-score distribution, revealing instances of attempted cheating for inflated scores and identifying students with language proficiency but limited motivation for coursework. This paper introduces a unique approach to investigating student learning processes, yielding insightful outcomes.

Background For automated analysis of metaphase chromosome images, the chromosome objects on the images need to be segmented in advance. However, the segmentation results often contain a lot of non-chromosome objects in the images. Hence, elimination of non-chromosome objects is an essential process in automated chromosome image analysis. This study aims to exclude non-chromosome objects and preserve as many chromosomes as possible. In this paper, we propose a hybrid deep learning method to exclude non-chromosome objects from metaphase chromosome images. Method The proposed method consists of two phases. In the first phase, two classification results are obtained from feature-based and image-based convolutional neural networks (CNN) separately (the feature-based CNN uses the features of the images as input; the image-based CNN uses the images as input directly). In the second phase, the prediction results from the above two CNNs are combined and resent to another CNN to obtain final classification results. Results The proposed method uses 18,757 non-chromosome objects and 43,398 chromosomes (including single and multiple overlapped chromosomes) from 1038 chromosome images to evaluate the performance. The experimental results show that the proposed method can detect and exclude 99.61% (18,683/18,757) of the non-chromosome objects and preserve 99.95% (43,375/43,398) of the chromosomes for further analysis. Conclusions The proposed method has a high effectiveness on excluding non-chromosome objects and could be used as a preprocessing procedure for chromosome image analysis.