The loss function is a key component in deep learning models. A commonly used loss function for classification is the cross entropy loss, which is a simple yet effective application of information theory for classification problems. Based on this loss, many other loss functions have been proposed, e.g., by adding intra-class and inter-class constraints to enhance the discriminative ability of the learned features. However, these loss functions fail to consider the connections between the feature distribution and the model structure. Aiming at addressing this problem, we propose a channel correlation loss (CC-Loss) that is able to constrain the specific relations between classes and channels as well as maintain the intra-class and the inter-class separability. CC-Loss uses a channel attention module to generate channel attention of features for each sample in the training stage. Next, an Euclidean distance matrix is calculated to make the channel attention vectors associated with the same class become identical and to increase the difference between different classes. Finally, we obtain a feature embedding with good intra-class compactness and inter-class separability. Experimental results show that two different backbone models trained with the proposed CC-Loss outperform the state-of-the-art loss functions on three image classification datasets. © 2020 IEEE

Face recognition usually uses different features as input signals. There are many conventional techniques being used and these include technical and fundamental analysis. In this paper, the sample data is mapped from low-dimensional space to high-dimensional space by the kernel method, which makes the classification algorithm have the ability to deal with non-linear data and can solve the small sample problem. At the same time, deep belief network is used as feature transformation and classification to mine feature information of high-dimensional face data. The experimental results show that the optimal recognition rate of the proposed algorithm in a specific face database is up to 96%.
© Published under licence by IOP Publishing Ltd.

In the process of learning with the help of online platform, students' learning satisfaction is an important factor that constitutes the effect of online teaching. On the basis of the existing research, this study proposes three factors, namely, cognitive level, technological environment and social environment, to constitute online learning satisfaction. At the same time, through statistical analysis, the key factors affecting learning satisfaction are sorted out, the selection model of identifying the key factors affecting learning satisfaction is established, and online learning satisfaction is fitted by prediction. The study shows that self-efficacy evaluation, teaching support evaluation, platform use evaluation and student-teacher interaction evaluation have a great impact on students' learning satisfaction in the process of online learning. At the same time, the 93% prediction accuracy of the prediction model based on the above-mentioned research can be achieved. In addition, this paper also puts forward suggestions on how to strengthen and improve learning satisfaction.
© Published under licence by IOP Publishing Ltd.

In order to predict the short-term load variation range and trend of cloud computing, this paper proposed a prediction model based on information granulation support vector machine (IGSVM). Taking the historical load value as a sample to do simulation training, through Gravitational Search Algorithm (GSA) to optimize the parameters of SVM, and make regression prediction to three parameters of triangular fuzzy particles, Low, R and Up, to obtain the variation range and trend of short-term load. The result is consistent with the actual situation, which verifies the validity of the model and provides the basis for actual operation and maintenance.
© Published under licence by IOP Publishing Ltd.