International Journal of Computational Intelligence and Applications

Universities play a huge role in the cultivation of talents. Especially in the context of internationalization, the teaching of English as a common language is becoming more and more important. This paper introduced the traditional methods for evaluating the quality of English teaching, established a deep learning algorithm for evaluating the quality of English teaching with the evaluation indicators of the traditional methods combined with the convolutional neural network (CNN) algorithm, conducted simulation experiments on the CNN algorithm, and compared it with the support vector machine (SVM) algorithm. The results showed that the scores obtained by the CNN algorithm had some errors with the actual scores but were much lower than the scores obtained by the SVM algorithm, and the CNN algorithm consumed a shorter time in computing. This paper used the CNN algorithm combined with evaluation indexes constructed by the analytic hierarchy process (AHP) method to evaluate the quality of English teaching and verified the effectiveness of the CNN algorithm through a comparison with the SVM algorithm, which provides an effective reference for intelligent evaluation of English teaching quality.

This paper studied music feature recognition and classification. First, the common signal features were analyzed, and the signal pre-processing method was introduced. Then, the Mel–Phon coefficient (MPC) was proposed as a feature for subsequent recognition and classification. The deep belief network (DBN) model was applied and improved by the gray wolf optimization (GWO) algorithm to get the GWO–DBN model. The experiments were conducted on GTZAN and free music archive (FMA) datasets. It was found that the best hidden-layer structure of DBN was 1440-960-480-300. Compared with machine learning methods such as decision trees, the DBN model had better classification performance in recognizing and classifying music types. The classification accuracy of the GWO–DBN model reached 75.67%. The experimental results demonstrate the reliability of the GWO–DBN model. The GWO–DBN model can be further promoted and applied in actual music research.

In the modern capital market, the price of a stock is often considered to be highly volatile and unpredictable because of various social, financial, political and other dynamic factors. With calculated and thoughtful investment, stock market can ensure a handsome profit with minimal capital investment, while incorrect prediction can easily bring catastrophic financial loss to the investors. This paper introduces the application of a recently introduced machine learning model — the transformer model, to predict the future price of stocks of Dhaka Stock Exchange (DSE), the leading stock exchange in Bangladesh. The transformer model has been widely leveraged for natural language processing and computer vision tasks, but, to the best of our knowledge, has never been used for stock price prediction task task using DSE data. Recently, the introduction of time2vec encoding to represent the time series features has made it possible to employ the transformer model for the stock price prediction. This paper aims to leverage these two effective techniques to discover forecasting ability on the volatile stock market of DSE. We deal with the historical daily and weekly data of eight specific stocks listed in DSE. Our experiments demonstrate promising results and acceptable root-mean-squared error on most of the stocks. We also compare the performance of our model with that of a well-known benchmark stock forecasting model called ARIMA and report satisfactory results.

In order to detect and identify fire accidents accurately and efficiently, an intelligent fire identification system based on neural network algorithm is designed, which can overcome the shortcomings of single information, complex wiring, poor adaptability, etc. The characteristic extraction of sensors is adopted in the information layer to solve the problems in multi-sensor fusion. The fire data are transmitted to the main controller through LoRa wireless module and fused by back propagation neural network, which is self-learning and adaptive. The output of neural network and fuzzy inference with other factors are used for decision criteria to improve the identification accuracy. The common combustibles and various interference sources are selected for fire tests. The result shows that the detection accuracy is up to 100% and the false alarm rate is lower than 0.1%, meanwhile, the system has the advantages of fast response and high detection efficiency.

How to utilize symbolic knowledge in deep learning is an important problem. Deep neural networks are flexible and powerful, while symbolic knowledge has the virtue of interpretability and intuitiveness. It is necessary to combine the two together to inject symbolic knowledge into neural networks. We propose a novel approach to combine neural networks with logic rules. In this approach, task-specific supervised learning and policy-based reinforcement learning are performed alternately to train a neural model until convergence. The basic idea is to use supervised learning to train a deep model and use reinforcement learning to propel the deep model to meet logic rules. In the process of the policy gradient reinforcement learning, if a predicted output of a deep model meets all logical rules, the deep model is given a positive reward, otherwise, it is given a negative reward. By maximizing the expected rewards, the deep model can be gradually adjusted to meet logical constraints. We conduct experiments on the tasks of named entity recognition. The experimental results demonstrate the effectiveness of our method.

In recent years, with the strong and outstanding development of the Internet, the need to refer to the feedback of previous customers when shopping online is increasing. Therefore, websites are developed to allow users to share experiences, reviews, comments and feedback about the services and products of businesses and organizations. The organizations also collect user feedback about their products and services to give better directions. However, with a large amount of user feedback about certain services and products, it is difficult for users, businesses, and organizations to pay attention to them all. Thus, an automatic system is necessary to analyze the sentiment of a customer feedback. Recently, the well-known pre-trained language models for Vietnamese (PhoBERT) achieved high performance in comparison with other approaches. However, this method may not focus on the local information in the text like phrases or fragments. In this paper, we propose a Convolutional Neural Network (CNN) model based on PhoBERT for sentiment classification. The output of contextualized embeddings of the PhoBERT’s last four layers is fed into the CNN. This makes the network capable of obtaining more local information from the sentiment. Besides, the PhoBERT output is also given to the transformer encoder layers in order to employ the self-attention technique, and this also makes the model more focused on the important information of the sentiment segments. The experimental results demonstrate that the proposed approach gives competitive performance compared to the existing studies on three public datasets with the opinions of Vietnamese people.

Nowadays, images have become one of the most popular forms of communication as image editing tools have evolved. Image manipulation, particularly image colorization, has become easier, making it harder to differentiate between fake colorized images and actual images. Furthermore, the RGB space is no longer considered to be the best option for color-based detection techniques due to the high correlation between channels and its blending of luminance and chrominance information. This paper proposes a new approach for fake colorized image detection based on a novel image representation created by combining color information from three separate color spaces (HSV, Lab, and Ycbcr) and selecting the most different channels from each color space to reconstruct the image. Features from the proposed image representation are extracted based on transfer learning using the pre-trained CNNs ResNet50 model. The Support Vector Machine (SVM) approach has been used for classification purposes due to its high ability for generalization. Our experiments indicate that our proposed models outperform other state-of-the-art fake colorized image detection methods in several aspects.

Few-shot learning can solve new learning tasks in the condition of fewer samples. However, currently, the few-shot learning algorithms mostly use the ResNet as a backbone, which leads to a large number of model parameters. To deal with the problem, a lightweight backbone named DenseAttentionNet which is based on the Convolutional Block Attention Mechanism is proposed by comparing the parameter amount and the accuracy of few-shot classification with ResNet-12. Then, based on the DenseAttentionNet, a few-shot learning algorithm called Meta-DenseAttention is presented to balance the model parameters and the classification effect. The dense connection and attention mechanism are combined to meet the requirements of fewer parameters and to achieve a good classification effect for the first time. The experimental results show that the DenseAttentionNet, not only reduces the number of parameters by 55% but also outperforms other classic backbones in the classification effect compared with the ResNet-12 benchmark. In addition, Meta-DenseAttention has an accuracy of 56.57% (5way-1shot) and 72.73% (5way-5shot) on the miniImageNet, although the number of parameters is only 3.6M. The experimental results also show that the few-shot learning algorithm proposed in this paper not only guarantees classification accuracy but also has the characteristics of lightweight.