International Journal of Engineering Intelligent Systems

Innovation of College English Teaching Methods Based on Deep Learning

Gangtao Chen, Qi Zhang, Yuqian Han — Tue, 01 Jul 2025 00:00:00 -0700

To provide innovative strategies to ensure the richness of teaching content, flexibility of teaching methods and efficiency of teaching process, this study uses an empirical experiment to explore the effectiveness of a college English teaching model based on deep learning. Two groups of 150 college students were selected as experimental subjects. One group received a personalized teaching mode based on deep learning, and the control group were taught by means of traditional pedagogy. The study compared the differences between the two groups in terms of English proficiency, learning attitudes and habits, learning efficiency, affective cognition and collaborative teamwork. The results show that the experimental group was significantly
better than the control group according to all the evaluation indicators, indicating that the teaching model based on deep learning has significant advantages in improving the effectiveness of students’ learning, improving learning attitudes and habits, improving learning efficiency, enhancing emotional cognitive stability and strengthening collaborative and cooperative team skills. These findings provide strong evidence for the improvement and optimization of college English teaching methods, and prove that the teaching model based on deep learning has great potential in promoting college students’ English skills in the reading, writing and speaking of English.

Keywords: deep learning; college English; English teaching; innovative method

Cite As

G. Chen, Q. Zhang, Y. Han, " Innovation of College English Teaching Methods Based on Deep Learning",
Engineering Intelligent Systems, vol. 33 no. 4, pp. 355-364, 2025.

Optimizing Quantitative Intelligent Systems Investment Strategies and Asset Allocation with Machine Learning

Zhanyong Wu, Xiaohang Ma, Yanxue Li — Tue, 01 Jul 2025 00:00:00 -0700

In the decision-making process of financial markets, machine learning techniques, especially support vector machines, have shown their potential to improve the efficiency and accuracy of investment strategies. This study explores the application of machine learning for the optimization of quantitative investment strategies and asset allocation by constructing and optimizing an SVM-based multi-factor stock selection model and asset allocation system. This study verifies the actual performance of the proposed SVM model in the financial market and its ability in terms of risk control and maximization of returns. The results show that SVM provides a higher rate of return and lower risk than traditional investment methods, which confirms its application value in modern financial strategies. This study provides a new perspective on, and technical support for, the field of quantitative investment and presents theoretical support for the further development and application of machine learning technology in the financial market.

Keywords: machine learning; support vector machine; finance; investment strategy; asset allocation

Cite As

Z. Wu, X. Ma, Y. Li, " Optimizing Quantitative Intelligent Systems Investment Strategies and Asset Allocation
with Machine Learning", Engineering Intelligent Systems, vol. 33 no. 4, pp. 365-375, 2025.

Algorithm for Detection and Defense of Neural Network Technology Based on Neural Network and Multimedia

Hui Ke — Tue, 01 Jul 2025 00:00:00 -0700

Neural networks have been widely usedin thefields ofimage recognition, voice recognition, and natural language processing among others. However, the neural network model is less robust against adversarial samples. That is, when small perturbations are artificially added in the input data, the output of the model will change. This phenomenon, known as adversarial example attack, can lead to misclassification and performance degradation of the model. In response to this problem, the academic community has proposed a variety of adversarial sample attack detection and defense methods. Adversarial attack detection is intended to detect adversarial samples and filter them out so that they are not input into the model. On the other hand, the purpose of adversarial defense is to increase the robustness of the model during training, making it more stable against adversarial samples. At present, there are still some problems in regard to the detection and defense of adversarial sample attacks. Therefore, further research and exploration are still of great significance. In this study, we examine “Adversarial Sample Attack and Defense in Neural Networks” to determine the protection capabilities of neural network technology. Through the experiment, the detection experiment of adversarial samples is carried out, and the three attack methods of C&W, FGSM and FIA are detected by using the adversarial detection algorithm based on neural network technology, and the detection success rate is recorded. Experimental results show that the average detection success rate of C&W, FGSM and FIA using the adversarial
detection algorithm based on neural network technology is 97.257%, 95.354% and 94.602%, respectively. This indicates that the algorithm has a high detection success rate for these three attack methods, can effectively identify adversarial samples, and improve the robustness and accuracy of the model.

Keywords: system attack, neural network, against the samples, image recognition

Cite As

H. Ke, "Algorithm for Detection and Defense of Neural Network Technology Based on Neural Network
and Multimedia", Engineering Intelligent Systems, vol. 33 no. 4, pp. 377-384, 2025.

Eye Tracking and Gesture Intelligent Systems Recognition System for Intelligent Human-Computer Interaction

Ke Wang, Weihua Zhu — Tue, 01 Jul 2025 00:00:00 -0700

In response to the issue of decreased accuracy of eye tracking and gesture recognition systems under different lighting conditions, user postures, and hand occlusions, the aim of this study is to enhance the human-computer interaction experience in smart home scenarios and construct an intelligent human-computer interaction system. The system captures eye movement images using infrared ray sources and cameras, processes iris edges using Canny edge detection, and captures changes in the user’s posture using the Lucas-Kanade optical flow algorithm to improve the accuracy of eye tracking. Multi-scale convolutional neural network (CNN) layers are designed, and a self-attention mechanism is applied to enhance the precision
of hand feature extraction. Long short-term memory (LSTM) networks are used to improve the accuracy of dynamic gesture recognition and achieve a more natural and intuitive interaction mode. The experimental results show that the accuracy of eye tracking reaches 95% at a light intensity of 1000 lux, and that of gesture recognition is the lowest at 89% among all light intensity conditions tested. In the case of partial occlusion, the highest success rate of gesture recognition is 95%. After interferences such as head movement and high-frequency blinking are added, the success rates of eye tracking are 95% and 94%, respectively. The experimental results demonstrate the efficiency of the system under good lighting conditions and its robustness in capturing complex gestures and occlusion situations, showing the accuracy of the optimized system for smart home control.

Keywords: Intelligent Human-Computer Interaction, Eye Tracking, Gesture Recognition, Convolutional Neural Networks, Long Short-Term Memory

Cite As

K. Wang, W. Zhu, "Eye Tracking and Gesture Intelligent Systems Recognition System for Intelligent Human-Computer
Interaction", Engineering Intelligent Systems, vol. 33 no. 4, pp. 385-395, 2025.

Implementation of Voice Software Testing Framework for Intelligent AI Technology

Chao Iv — Tue, 01 Jul 2025 00:00:00 -0700

The voice software testing framework is an information-based testing method aimed at verifying the functionality, reliability, and performance of voice applications, and plays a crucial role in the testing of various voice applications. The main role of testing frameworks is to provide high-quality and stable testing, and to help developers better understand the quality and performance of voice applications. Artificial intelligence (AI) technology plays an important role in voice software testing frameworks. By utilizing AI technology, automated testing can be achieved to improve test coverage and efficiency and identify more potential problems. In this project, the voice software testing framework was studied through AI technology. By conducting comparative experiments, the traditional Hidden Markov Model (HMM) andalgorithms based on AI technology werecompared to evaluate their performance. After experimental verification, it was found that in the field of voice recognition, the average recognition speeds of traditional HMMalgorithms and voice recognition judgment algorithms based on AI technology for Chinese, English, and emotional audio were 579ms, 568ms, 623ms, and 533ms, 526ms, and589ms, respectively. Experimental results showed that voice recognition judgment algorithms based on A technology performed better in terms of recognition speed for different audio types. In addition, this algorithm also outperformed traditional HMM algorithms in terms of recognition error rates for various audio types, indicating that it has stronger audio recognition capabilities. The excellent performance of voice recognition judgment algorithms based on AI technology has been demonstrated through experiments, providing a new direction for the design and research of voice software testing frameworks.

Keywords:Voice Software, Artificial Intelligence, Test Framework, Voice Recognition and Judgment Algorithm

Cite As

C. Iv, "Implementation of Voice Software Testing Framework for Intelligent AI Technology", Engineering Intelligent Systems, vol. 33 no. 4,
pp. 397-404, 2025.

Deep Learning Model for Image Classification in Machine Vision

Xiaojun Zhang — Tue, 01 Jul 2025 00:00:00 -0700

In this study, a deep-learning model based on Convolutional Neural Network (CNN) was applied in order to automatically extract image features and improve the accuracy and robustness of classification. A CNN model with multi-layer convolution and pooling structure was constructed, and the ReLU activation function was utilized to improve the ability of nonlinear expression. Multi-category classification was achieved through the Softmax layer. The training dataset was expanded through data augmentation technology to decrease the risk of overfitting and increase the adaptability of the model in complex scenes. A transfer learning strategy was adopted to use the model pre-trained on the large-scale dataset, ImageNet, to decrease the dependence on labeled data, and the model weights were optimized by fine-tuning. Using depthwise separable convolution, a lightweight network
structure was implemented to optimize the resource constraints of edge devices. Experimental results showed that the average inference latency of the model was 50.76ms and the average throughput was 20.2 FPS. The model still maintained high classification performance in a low computing environment.

Keywords: Deep Learning; Convolutional Neural Network; Data Augmentation; Transfer Learning; Lightweight Model

Cite As

X. Zhang, "Deep Learning Model for Image Classification in Machine Vision", Engineering Intelligent Systems,
vol. 33 no. 4, pp. 405-415, 2025.

Address Query and Signal Processing System Based Intelligent Systems on Artificial Intelligence Sensor Array in Time Domain

Haixia Guo — Tue, 01 Jul 2025 00:00:00 -0700

With the continuous development of the national economy and national defense technology, array address query signal processing systems are increasingly being applied in various fields such as communication, radar, seismic exploration, radio astronomy, wireless communication signal monitoring and management, mobile communication, navigation, medicine, etc. However, when multiple signals with the same beamwidth appear simultaneously, traditional array-based signal processing methods cannot effectively distinguish multiple signals. Artificial intelligence (AI) sensor array can address query and signal processing, and is an emerging information processing mode that breaks through the Rayleigh limitation of traditional
methods, thereby achieving target analysis and localization. This study applied the array time-domain technology in AI sensors to address query and signal processing systems. For the construction of an array address query and signal processing system, multiple signal classification algorithms were used to calculate signal fluctuations, and four data features of the two were compared and analyzed. The results showed that the proposed intelligent array address query and signal processing system can solve the problems existing in traditional methods, and the algorithm used can greatly improve the resolution ability of the array antenna and ease the restrictions on column structure. The time-domain address query and signal processing system of AI sensor arrays can maintain an average stability of over 86%, with a minimum time requirement of 32 seconds; accuracy generally consistent at over 90%, and an average number of 51.8 more than traditional systems in terms of universality. This plays a very important role in the research and development of both civilian and military equipment, and can facilitate further upgrading and development of the respective technology.

Keywords: Address Signal Processing System, Artificial Intelligence Sensor, Array Time Domain, Multiple Signal Classification Algorithm

Cite As

H. Guo, "Address Query and Signal Processing System Based Intelligent Systems on Artificial Intelligence Sensor Array in Time Domain", Engineering Intelligent Systems, vol. 33 no. 4, pp. 417-426, 2025.

A Power Data Sharing Scheme Based on Access Policy and Rewritable Blockchain–Access Policies and Power Data Sharing for Rewritable Blockchains

Shu Yang, Ziqiang Zhou, Wen Chai, Huanyu Wang — Tue, 01 Jul 2025 00:00:00 -0700

With the rapid development of smart grid technology, data security issues are gradually emerging, especially in terms of confidentiality. Although traditional encryption methods can protect the security of data, they often result in slow processing speed due to low encryption and decryption efficiency, and data is easily tampered with during transmission, posing a serious threat to the security and reliability of power data and the power system. In response to these issues, this article proposes an efficient power data sharing scheme based on Access Control Policy for Rewritable Blockchain Based onAccess Control Policy (RBACP).The immutability of blockchain makes power data more secure and transparent. The introduction of access control policies allows only those users who meet the permissions to make changes to data. This article implements an attribute-based
encryption This article implements an attribute-based encryption scheme for access control using identity-based signature embedding, resulting in attribute-based encryption for verifiable signatures (ABEVS), which strengthens the verification process, further avoids tampering, and improves data integrity. Finally, this article presents a specific example of RBACP and verifies its security and practicality through experiments and evaluations.

Keywords: smart grid; attribute-based encryption; rewritable blockchain; access strategy

Cite As

S. Yang, Z. Zhou, W. Chai, H. Wang, "A Power Data Sharing Scheme Based on Access Policy
and Rewritable Blockchain–Access Policies and Power Data Sharing for Rewritable Blockchains",
Engineering Intelligent Systems, vol. 33 no. 4, pp. 427-434, 2025.

Aseismic Optimization Design of Building Structures Based on Practical Reliability Algorithm

Hongsheng Yu, Juan Yin, Xiaoli Ren, Chengbin Xing, Youping Ding — Tue, 01 Jul 2025 00:00:00 -0700

Construction technology is closely related to social and economic development and directly affects personal safety. With the development of science and technology and the gradual improvement of architectural aesthetic and functional requirements, the height, safety and functionality of buildings have been significantly improved. In particular, the aseismic design of building structures plays a key role in improving architectural safety. In addition, the application of new materials and the improvement of construction management technology make aseismic design more applicable. Moreover, the structural reliability can objectively consider the influence and role of various random phenomena in the process of planning, construction and use of engineering structures, so as to optimize the balance between the safety and economy of building structures. At present, many high-rise buildings cannot effectively cope with the vibration and impact caused by earthquakes. The main reason is that the structural quality of high-rise buildings cannot respond well to large earthquakes. Therefore, this study analyzed the aseismic design of the building structure using the reliability practical algorithm, and optimized it so as to improve the aseismic ability of the building structure. The results showed that the aseismic capacity and structural load of the optimized building structure design were higher than the traditional ones; the aseismic capacity was about 9% higher, and the structural load was 6% higher.

Keywords: aseismic design of building structure, practical algorithm of reliability, aseismic structure optimization, optimal design

Cite As

H. Yu, J. Yin, X. Ren, C. Xing, Y. Ding, " Aseismic Optimization Design of Building Structures Based on Practical Reliability Algorithm", Engineering Intelligent Systems, vol. 33 no. 4, pp. 435-443, 2025.

Mining of Potential Fields and Structure Optimization of ESI Discipline in Universities Based on Artificial Intelligence Algorithm

Fang Zhao, Yuan Tu — Tue, 01 Jul 2025 00:00:00 -0700

This study examines the application of artificial intelligence (AI) algorithms in optimizing discipline layout and identifying potential fields in higher education. The objective is to enhance the development potential and resource allocation efficiency of essential science indicators (ESI) disciplines in universities through technical means. The paper explores the theoretical foundations and ESI discipline evaluation criteria based on artificial intelligence technology, providing an in-depth analysis of these technologies’ specific applications in discipline layout optimization. By means of comprehensive data collection and preprocessing, this study establishes a framework for algorithm design and performance evaluation, enabling the precise identification and analysis of potential fields for discipline development. Furthermore, the practical utility of the algorithm in optimizing discipline layout strategies is examined and discussed, offering suggestions for algorithm improvement and future development directions. This work provides a scientific basis for optimal decision-making regarding courses layout in colleges and universities, supporting the optimal allocation of higher education resources and the effective formulation of discipline development strategies.

Keywords: Artificial Intelligence algorithm, ESI subject evaluation, Optimization of discipline layout, Potential field mining

Cite As

F. Zhao, Y. Tu, "Mining of Potential Fields and Structure Optimization of ESI Discipline in Universities Based on Artificial Intelligence Algorithm", Engineering Intelligent Systems, vol. 33 no. 4, pp. 445-455, 2025.

Key Technologies for Financial Data Security Protection Based on Blockchain

Su Meng — Tue, 01 Jul 2025 00:00:00 -0700

Thecentralized database structure ofthe traditional financial system hassignificant risks in terms ofaudit and internal control, increasing its vulnerability to hacker attacks and making it difficult to trace and audit data. In addition, data privacy protection is inadequate and there is a high risk of information leakage. This study used the decentralized and tamper-proof characteristics of blockchain technology (BT) to improve the security and credibility of financial data in audit and internal control. The research used distributed storage to ensure data anti-attack capabilities, and smart contracts automatically perform data verification and management operations to reduce the impact of human intervention on the internal control process. Data encryption is combined with zero-knowledge proof methods to achieve privacy protection. Timestamp-based traceability and audit mechanisms are designed to ensure financial data compliance and verifiability. In terms of data consistency between nodes, the PBFT (practical Byzantian Fault Tolerance) consensus algorithm is used to optimize the efficiency of financial information synchronization, and the accurate access of authorized users in financial and audit activities is guaranteed through the role-based authority control system. Experimental results show that the blockchain based solution has a successful interception rate of more than 90%, and a data integrity verification pass rate of 100%, while the interception rate of the traditional DES (Data Encryption Standard) encryption solution does not exceed 70%. In regard to data privacy protection, the leakage rate of the zero-knowledge proof mechanism is only 0.35%, which is better than the 3% of RSA (Rivest-Shamir-Adleman) encryption. For financial data
tracing and auditing, the average tracing time of BT is 0.46 seconds, which is significantly better than the 2.22 seconds of traditional databases. The comprehensive evaluation results show that BT has excellent anti-attack ability, data privacy protection performance and financial data management efficiency in audit and internal control, and is expected to play an important role in a wider range of financial management applications.

Keywords: Financial Data Security Protection, Blockchain Technology, Data Privacy Protection, Smart Contracts, Attack Resistance Capability

Cite As

S. Meng, "Key Technologies for Financial Data Security Protection Based on Blockchain", Engineering Intelligent Systems, vol. 33
no. 4, pp. 457-467, 2025.

Decision Support System for Human Resource Management Based on Big Data

Ting He, Tingting Liu — Tue, 01 Jul 2025 00:00:00 -0700

This paper discusses the importance of building a human resource management decision support system based on big data technology, and analyzes its application value and challenges in human resource management (HRM). By reviewing the research progress of relevant domestic and foreign literature, a new design framework is proposed. In this study, a decision support system that integrates the internal and external big data of an enterprise is designed and implemented to improve the quality and efficiency of HRM decisions. In order to verify the effectiveness of the system, the study conducted empirical comparisons in scenarios such as talent selection, training needs analysis, and salary and benefits strategy formulation. The experimental results show that the new system improves the accuracy, recall rate, and F1 value of talent selection by 20.00%, 25.71%, and 13.25% respectively, and increases the correctness of decision making by 26.47%, improves decision efficiency by 50%, reduces recruitment costs by 37.5%, and increases recruitment satisfaction by 16.67%. In terms of training needs analysis, the accuracy, recall rate, F1 value, correct decision rate and user satisfaction of the new system increased by 21.43%, 27.69%, 25.37% and 28.13% respectively, and the decision-making efficiency increased by 50%. The training effect score and satisfaction increased by 17.14% and 23.53% respectively, and the training cost decreased by 33.33%. In terms of salary and benefits strategy formulation, the new system also made significant progress, with an accuracy increase of 20.55%, a recall increase of 26.47%, an F1 value increase of 24.29%, a decision correctness increase of 27.27%, a decision efficiency increase of 50%, and a salary satisfaction score increase of 38.71%. The innovative contribution of the research is that the system optimizes the human resource allocation process through
intelligent means, significantly improves the efficiency and effectiveness of various HRM tasks, and achieves the in-depth mining and value creation of the human resource data of enterprises.

Keywords: big data, human resources, management decisions, decision support systems

Cite As

T. He, T. Liu, "Decision Support System for Human Resource Management Based on Big Data", Engineering Intelligent Systems, vol. 33 no. 4, pp. 469-480, 2025.