Applied and Computational Engineering

- The Open Access Proceedings Series for Conferences

Volume Info.

  • Title

    Proceedings of the 2023 International Conference on Machine Learning and Automation

    Conference Date

    2023-10-18

    Website

    https://2023.confmla.org/

    Notes

     

    ISBN

    978-1-83558-303-6 (Print)

    978-1-83558-304-3 (Online)

    Published Date

    2024-02-21

    Editors

    Mustafa İSTANBULLU, Cukurova University

Articles

  • Open Access | Article 2024-02-21 Doi: 10.54254/2755-2721/39/20230566

    McDonald's food target recognition and calorie display based on YOLOv5 algorithm

    This research paper delves into the development and assessment of a novel food recognition and evaluation system tailored for McDonald's menu items, leveraging the capabilities of the YOLOv5 algorithm. The study demonstrates that the system can successfully identify McDonald's food items from images and seamlessly query calorie and nutritional information from a backend database. The data is then presented to the user, aiding in more informed dietary choices and promoting public health awareness. The system has particular utility for McDonald's customers, facilitating real-time decisions that align with individual health goals and nutritional requirements. Our experimental findings show a high degree of accuracy and efficiency, although the system's scope is currently limited to five key menu items. Future directions for this work include expanding the range of recognizable food categories and implementing user feedback mechanisms to refine recognition accuracy. Moreover, the paper discusses potential optimizations for reducing system response time and further enhancing the practical utility of the technology. This research serves as a significant step towards utilizing computer vision technologies for public health interventions, aiming to combat the rise of obesity and related diseases.

  • Open Access | Article 2024-02-21 Doi: 10.54254/2755-2721/39/20230570

    Research on autonomous mobile robot maze navigation problem based on Dijkstra’s algorithm

    In recent years, the field of autonomous mobile robotics has garnered significant attention due to its potential applications in various domains such as logistics, surveillance, and search and rescue operations. A crucial challenge in this area is the efficient navigation of robots within complex and dynamic environments, particularly when navigating through maze-like structures. The maze navigation problem involves finding optimal paths for robots to traverse from their initial positions to designated destinations while avoiding obstacles and making intelligent decisions to ensure timely and safe navigation. This study aims to investigate and apply Dijkstra’s algorithm to solve the maze navigation problem for autonomous mobile robots. By analyzing the navigation challenges faced by autonomous mobile robots in maze environments, a solution based on Dijkstra’s algorithm is proposed. In conclusion, this study contributes to the field of autonomous mobile robotics by proposing and evaluating the application of Dijkstra’s algorithm for maze navigation. The experimental results validate its potential to address the challenges of navigating intricate maze environments. However, it is acknowledged that further refinement and innovation are possible to continue improving the performance of autonomous mobile robots in maze navigation scenarios.

  • Open Access | Article 2024-02-21 Doi: 10.54254/2755-2721/39/20230571

    Guiding to available parking spots with AR device

    Parking congestion has become a pressing issue in urban areas, leading to increased traffic congestion and environmental pollution. This paper presents a comprehensive study on intelligent parking systems aimed at addressing these challenges. The study targets urban drivers, transport authorities and citizens who are affected by parking difficulties and traffic congestion. Intelligent Parking System is a technology-driven solution that aims to optimise parking space utilisation and reduce traffic congestion. It efficiently guides drivers to available parking spaces through real-time data collection and analysis. The study was conducted in Unity3D environment over a period of 2 weeks and covered the phases of system design, development, testing and implementation. The main motivation for this research was to mitigate the negative effects of parking congestion. Intelligent parking systems have the potential to increase traffic efficiency, reduce air pollution and improve the overall quality of life in cities. Server and Unity models were used in the study. Real-time sensors collect parking occupancy data, which is processed to provide real-time parking availability information to drivers via the server. In conclusion, this paper presents an intelligent parking system that can solve the challenges of parking congestion. By utilising server technology and data-driven insights, the project aims to improve the way cities manage parking and contribute to more sustainable urban mobility.

  • Open Access | Article 2024-02-21 Doi: 10.54254/2755-2721/39/20230573

    An evaluation of the impact of ChatGPT on network security

    ChatGPT, a very advanced natural language generation model, represents a momentous paradigm shift inside the realm of the internet. ChatGPT, which was made available to the public by OpenAI on November 30, 2022, is an enhanced version of OpenAI’s GPT-3.5 model. It has been further developed through the implementation of fine-tuning methods that combine both supervised and reinforcement learning approaches. Furthermore, it provides a client interface that is easy to use, allowing users to actively participate in interactive question-and-answer exchanges with the model. Nevertheless, the utilization of these chatbots likewise presents noteworthy cybersecurity concerns that necessitate attention. The primary objective of this research study is to examine the cyber dangers that are inherent in the utilization of ChatGPT and other comparable AI-driven chatbots. This investigation will encompass an analysis of potential vulnerabilities that may be susceptible to exploitation by individuals with malevolent intent. Additionally, the paper proposes strategies for mitigating the aforementioned cyber risks and vulnerabilities.

  • Open Access | Article 2024-02-21 Doi: 10.54254/2755-2721/39/20230574

    Using data resampling and category weight adjustment to solve sample imbalance

    The purpose of this thesis is to investigate the application of artificial intelligence and machine learning in solving the sample imbalance problem. The sample imbalance problem refers to the phenomenon that the number of different categories of samples in the training data varies greatly, resulting in the poor performance of traditional machine learning algorithms on a few categories of samples. To address this problem, this paper proposes a new approach combining data resampling and category weight adjustment strategies. First, the sample distribution of the dataset is adjusted by undersampling and oversampling techniques to balance the number of samples from different categories. Then, during the model training process, different weights are assigned to the samples of different categories so that the model pays more attention to the samples of a few categories. The experimental results show that the method achieves significant performance improvement on multiple datasets. In addition, this paper compares other commonly used methods for solving the sample imbalance problem and analyzes and discusses them in detail. Finally, this study offers a practical solution to the problem of sample imbalance and provides guidance for research in related fields.

  • Open Access | Article 2024-02-21 Doi: 10.54254/2755-2721/39/20230575

    Deep learning based multi-target detection for roads

    The vehicle target detection algorithm based on deep learning has gradually become a research hotspot in this field. In recent years, with the significant breakthrough of deep learning in the field of target recognition, the vehicle target detection algorithm based on deep learning has gradually become a research hotspot in this field. For the task of vehicle target detection, this paper first briefly introduces the process of traditional target detection algorithms and some optimization methods. It summarizes the development process of YOLO, the current mainstream one-stage vehicle target detection algorithm, and the process of Faster R-CNN, the second-stage vehicle target detection algorithm, and its improvement. Then the characteristics of several types of representative convolutional neural network algorithms are analyzed in chronological development order. Finally, it looks forward to t he future research direction of vehicle target detection algorithms, and also provides new ideas for the optimization of the subsequent vehicle target detection algorithms, which have good engineering application value. Provides algorithmic support for the underlying logic of autonomous driving.

  • Open Access | Article 2024-02-21 Doi: 10.54254/2755-2721/39/20230576

    Further exploration on deep learning in flower recognition

    Flower recognition is an important research direction in the field of computer vision, and automatic classification of flower images through deep learning methods is of great significance for ecological environment monitoring and plant research. With this background, this study aims to further optimize the existing flower recognition system and improve its classification accuracy by adjusting key parameters. Based on the existing deep learning model and several rounds of training, this paper explores the tuning strategies for parameters such as different learning rates and weight decay to achieve higher recognition accuracy. In the experiments, this paper uses the classical flower dataset and enhances the diversity of the data through image preprocessing and data enhancement. Through a hundred rounds of training, the model in this paper achieves about 80% classification accuracy on the test set, which is significantly improved compared with the initial model. Further analysis of the results shows that by reasonably adjusting the learning rate and weight decay parameters, the model achieves certain improvements in different flower classes, demonstrating the impact of parameter tuning on the overall performance of the model.

  • Open Access | Article 2024-02-21 Doi: 10.54254/2755-2721/39/20230577

    Design and analysis of adaptive control systems: Adaptive control algorithms using machine learning

    A growing field of research is the use of adaptive control algorithms with machine learning techniques like Q-learning and SARSA. With prospective applications in robotics, healthcare, and other fields, this interdisciplinary method strives to combine the stability and robustness of conventional control systems with the self-learning skills of reinforcement learning. Making sure that systems are stable and that learning occurs effectively is the main research problem. This paper demonstrates the stability and convergence of existing adaptive control algorithms when integrating with machine learning. There are mainly four primary methods of control algorithms: reinforcement learning, neutral network, support vector machine and deep learning. Reinforcement learning is the main focus of this paper. Data efficiency, robustness, and generalization are the main problems with reinforcement learning. Q-learning and SARSA (State, Action, Reward, State’, Action’) are two algorithms for reinforcement learning. The research will be done by analyzing these two algorithms based on existing material and the actual application of these two algorithms. SARSA is believed to be more safe as its on-policy methodology, and Q-learning is believed to be more proactive as its off-policy methodology.

  • Open Access | Article 2024-02-21 Doi: 10.54254/2755-2721/39/20230579

    From ELIZA to ChatGPT: A brief history of chatbots and their evolution

    Over the years, chatbots have grown to be used in a variety of industries. From their humble beginnings to their current prominence, chatbots have come a long way. From the earliest chatbot ELIZA in the 1960s to today’s popular Chatgpt, chatbot language models, codes, and databases have improved greatly with the advancement of artificial intelligence technology.This paper introduces the development of chatbots through literature review and theoretical analysis. It also analyzes and summarizes the advantages and challenges of chatbots according to the current status of chatbot applications and social needs. Personalized interaction will be an important development direction for chatbots, because providing personalized responses through user data analysis can provide users with a personalized experience, thus increasing user engagement and satisfaction.

  • Open Access | Article 2024-02-21 Doi: 10.54254/2755-2721/39/20230581

    Research on multi-role classification task of online mall based on heterogeneous graph neural network

    With the rapid development of e-commerce, online shopping malls have become an indispensable part of daily life. In order to better meet the needs of consumers, marketplace platforms need to accurately identify and categorize different user personas to provide personalized services and recommendations. In traditional role classification methods, basic information and behavioral data of users are typically used for classification. However, this approach often ignores the complex relationships between users and multiple heterogeneous data such as goods, reviews, social networks, and more. Therefore, we propose a new approach based on heterogeneous graphs to model different types of data in the form of graphs to better capture the connections between users and various elements in the marketplace. In this study, graph embedding technology is used to map nodes in heterogeneous graphs into low-dimensional vector spaces to capture similarities and relationships between nodes. Then, using the vector representation of these nodes, we can apply algorithms such as attention mechanisms for multi-role classification. Specifically, we use algorithms such as support vector machines to train classification models and use heterogeneous graph attention mechanisms to obtain the final feature representation of nodes. Experimental results show that our method shows significant advantages in multi-role classification tasks. Finally, the results of this study are discussed and summarized. We found that the classification model based on heterogeneous graph can effectively classify multiple roles in the online mall to provide personalized services and recommendations for the mall. At the same time, we also find that the construction of heterogeneous maps and the choice of graph embedding technology have important impacts on the classification results, which need further research and optimization. Therefore, multi-role task classification of online shopping malls based on heterogeneous graph neural networks is of great significance for improving the user experience and recommendation effect of online shopping malls, and also provides new ideas and methods for research in related fields.

  • Open Access | Article 2024-02-21 Doi: 10.54254/2755-2721/39/20230582

    Exploring multiple algorithms for leaf classification: A comparative study

    In the world there are an estimated nearly half a million plant species, and the leaves of each plant have their own unique characteristics. However, leaf classification has historically been problematic, with some similar-looking leaves being repeatedly identified and errors may even occur. In the past, people used precision instruments, chemical analysis, and other methods to improve the success rate of resolution, but this often required a certain amount of professional knowledge, sometimes it was trouble to operate. And the development of computer in recent years makes the leaf classification problem can be solved better through the image processing and machine learning technology. Therefore, the topic of this paper is to classify leaves by different algorithms and compare the classification results, which include KNN, Random Forest, the neural network algorithm ANN, and CNN. It can be seen through experiments that the classification results obtained by ANN have a higher accuracy rate among these algorithms.

  • Open Access | Article 2024-02-21 Doi: 10.54254/2755-2721/39/20230583

    Sentiment analysis based on long short-term memory model

    An important area of natural language processing is text emotion analysis. Emotion analysis is a very meaningful research work and has a broad application prospect, such as social media monitoring, reputation research of commodity brands, market research, and so on. By analyzing the time and content factors of the data, considering the final application scenario, and finally comparing the advantages and disadvantages of the methods, There are three main techniques for analyzing emotions in text: emotion analysis using an emotion dictionary, emotion analysis using machine learning, and emotion analysis using deep learning. Among them, Convolutional neural networks (CNN), recurrent neural networks (RNN), and long short-term memory networks (LSTM) are examples of deep learning-based techniques. For processing temporal relate issues such as video, speech, and text, CNN algorithms often consume a large amount of computational time, especially for processing image datasets, which may encounter specific problems. To address this issue, RNN is more suitable for solving temporal related issues such as video, voice, and text. In natural language, word order is an extremely important feature. RNN may potentially process sequences of any length, add memory units based on the original neural network, handle pre- and post-word dependencies, and process sequences of any length. The main work is as follows: LSTM adds or deletes unit states through a structure called gate, Determine the experimental thinking of text analysis, Crawl and train data sets through AI Studio, pycharm and other tools.

  • Open Access | Article 2024-02-21 Doi: 10.54254/2755-2721/39/20230585

    A special target detection called You Only Looks Once

    This paper proposes a unique target detection called You Only Look Once (YOLO for short) and a recognition method. Unlike the previous idea of many classifiers with object detection functions, the object detection box is set as a spatially separated bounding box and the regression problem of related class probabilities is realized. The neural network model can directly scan the entire image during testing, predicting bounding boxes and class probabilities from the complete picture. At the same time, because the whole detection channel relies on a single neural network, it is more straightforward and concise when upgrading and updating. The unified architecture used in this paper is fast, with smaller versions of the YOLO model processing a staggering 155 frames per second. At the same time, it has also achieved excellent results in mAP. Compared with other detection systems, YOLOv3 has optimized many past problems, including positioning errors. At the same time, the probability of predicting false detections in the absence of false detections is small. Finally, like the YOLO base model, YOLOv3 may produce significant errors when processing abstract works of art and images with a large number of small objects. However, its actual results are still better than detection methods such as Region -Convolutional Neural Networks (R-CNN).

  • Open Access | Article 2024-02-21 Doi: 10.54254/2755-2721/39/20230586

    A comparison of recent progress in breast cancer diagnosis models using machine learning algorithms

    Breast cancer is the leading cause of mortality among women suffering from cancer, so the accurate diagnosis is important. This review aims to provide a thorough examination of advancements and trends in breast cancer diagnosis by analysing recognized papers published between 2020 and 2023.The paper firstly gives a brief overview of breast cancer, machine learning algorithms, followed by an introduction of basic process for ML in breast cancer diagnosis. After that, by focusing on two emerging trends, hybridization and newly invented modalities, the review introduces existing achievements in the field. Subsequently, it highlights nine notable or novel designs in breast cancer diagnosis, while presenting their comparative properties in a tabular format. Hopefully, this review can equip researchers with valuable insights for future studies and references, helping them gain a better understanding of the field and facilitating further improvements in breast cancer detection and classification.

  • Open Access | Article 2024-02-21 Doi: 10.54254/2755-2721/39/20230587

    Model optimization method based on the MOON algorithm

    Federated Learning is a revolutionary approach to machine learning. Its purpose is to enable multiple participants to collaboratively train machine learning models without the need to share local data. The main objective is to address issues related to data privacy and security. In traditional machine learning, data is typically centralized and stored in a single location or on cloud servers for training. However, this centralized training approach carries risks of potential data leakage, especially concerning sensitive and critical information. Industries such as healthcare and finance, which involve sensitive data, place a premium on safeguarding data privacy. Furthermore, in cases where data cannot be easily transferred or is subject to privacy regulations, centralized methods may face limitations. Federated Learning revolutionizes the conventional approach by introducing a decentralized model training process. It maintains data decentralization while achieving collaborative model optimization, greatly enhancing data privacy and security. The MOON algorithm, an integral part of federated learning, contributes to its novelty. As a significant component, the MOON algorithm facilitates new possibilities for federated learning. In this article, the research will elaborate on the MOON algorithm within the context of federated learning. And this article will delve into its description and optimization, elucidating how it enhances federated learning.

  • Open Access | Article 2024-02-21 Doi: 10.54254/2755-2721/39/20230588

    Asynchronous Federated Learning: Methods and applications

    Federated Learning (FL) is a distributed alternative to traditional machine learning frameworks that computes a global model on a centralized aggregation server according to the parameters of local models, which can address the privacy leakage problem caused by collecting sensitive data from local devices. However, the classic FL methods with synchronous aggregation strategies, in many cases, shall suffer from limitations in resource utilization due to the need to wait for slower devices (stragglers) to aggregate during each training epoch. In addition, the accuracy of the global model can be affected by the uneven distribution of data among unreliable devices in real-world scenarios. Therefore, many Asynchronous Federated Learning (AFL) methods have been developed on many occasions to improve communication efficiency, model performance, privacy, and security. This article elaborates on the existing research on AFL and its applications in many areas. The paper first introduces the concept and development of FL, and then discusses in detail the related work and main research directions of AFL, including dealing with stragglers, staleness, communication efficiency between devices, and privacy and scalability issues. Then, this paper also explores the application of AFL in different fields, especially in the fields of mobile device edge computing, Internet of Things devices, and medical data analysis. Finally, the article gives some outlook on future research directions and believes that it is necessary to design efficient asynchronous optimization algorithms, reduce communication overhead and computing resource usage, and explore new data privacy protection methods.

  • Open Access | Article 2024-02-21 Doi: 10.54254/2755-2721/39/20230589

    CNN-based image style transformation--Using VGG19

    eural Style Transfer is a widely used approach in the field of computer vision, which aims to generate visual effects by integrating the information contained in one image into another. In this paper, this work presents an implementation of neural style transfer using TensorFlow and the VGG19 model. The proposed method involves loading and preprocessing the content and style images, extracting features from both images using the VGG19 model, and computing Gram matrices to capture the style information. A StyleContentModel class is introduced to encapsulate the style and content extraction process. The optimization process is performed using the Adam optimizer, where gradients are applied to iteratively update the generated image. The number of epochs and steps per epoch can be adjusted to control the optimization process and achieve desired results. Experiments show that we are effective in generating an image that is able to integrate the content of one image into the other. The generated images exhibit visually appealing characteristics and showcase the potential of neural style transfer as a creative tool in image synthesis. Future work may involve exploring different variations of the style transfer algorithm, optimizing hyperparameters, and evaluating the performance on a wider range of image datasets. Additionally, the integration of other deep learning architectures and techniques could further enhance the capabilities of neural style transfer.

  • Open Access | Article 2024-02-21 Doi: 10.54254/2755-2721/39/20230591

    Federated learning method for dynamic weight adjustment: An in-depth analysis of FedDyn algorithm

    With the gradual enrichment of the application of deep learning in daily life, distributed machine learning has played a greater role in daily life, among which federated learning and its optimization algorithm FedDyn algorithm have begun to receive widespread attention. After introducing federated learning and traditional algorithms, this paper starts with the operation mode and framework of the FedDyn algorithm, takes the combination method of port selection strategy and global distribution as an example, analyzes the characteristics of some links in the operation process and uses the innovation points of the FedDyn algorithm : Dynamic weight adjustment is the entry point, which illustrates the advantages of the FedDyn algorithm in current use, and also points out some of the current problems. Finally, from the aspect of market application, the current application status of the algorithm is introduced from the aspects of personalized recommendation, medical treatment and finance, and the future development of the algorithm is prospected.

  • Open Access | Article 2024-02-21 Doi: 10.54254/2755-2721/39/20230592

    The experiment of federated learning algorithm

    As technology advances, concerns regarding data privacy and security have become prominent challenges in machine learning applications. Nevertheless, the introduction of federated learning technology has effectively tackled this concern by concurrently enhancing model performance and preserving data privacy, thereby presenting a more secure and efficient solution within our digital realm. Utilizing discussions about the background of federated learning technology, coupled with pertinent algorithmic procedures and logic, this paper proficiently implements the FedMA algorithm. Additionally, the study performs a comparative analysis of the accuracy and efficiency of the FedMA, FedAvg, FedDyn, and MOON algorithms utilizing the Fashion-MNIST dataset. Moreover, the investigation not only optimizes parameter tuning for the MOON algorithm but also extends experiments to the CIFAR-10 and AGNews datasets, thereby providing additional comparisons of performance and strengths among various federated learning algorithms. In conclusion, the paper provides a comprehensive summary and outlines potential avenues for future research. These insights enhance the comprehension of federated learning and offer valuable guidance to advance and refine its practical applications.

  • Open Access | Article 2024-02-21 Doi: 10.54254/2755-2721/39/20230593

    Face recognition technology based on ResNet-50

    Face recognition technology is progressively finding its place across diverse domains. In pursuit of enhancing the efficacy of face recognition systems, this study employs a ResNet-50 deep convolutional neural network. The dataset is meticulously gathered and processed via OpenCV, thus amplifying the precision and utility of face recognition. ResNet, an advanced convolutional neural network, incorporates the concept of residual connections, bridging convolutional layers through shortcut connections. These connections facilitate the addition of input to output, forming residual blocks. Consequently, ResNet-50 efficiently tackles the vanishing gradient issue, enabling the training of exceptionally deep networks. With 49 convolutional layers and a fully connected layer, ResNet-50 boasts a robust architecture. To emulate varying brightness conditions, post-collection image adjustments are applied randomly. This strategy curbs the impact of divergent lighting scenarios on recognition accuracy, bolstering the model’s practical applicability. Notably, experimental outcomes underscore the commendable performance of the trained ResNet-50 model in face recognition trials. This substantiates the broad-spectrum viability of face recognition technology in domains such as security surveillance, human-machine interaction, identity verification, and beyond.

Copyright © 2023 EWA Publishing. Unless Otherwise Stated