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

    978-1-83558-291-6 (Print)

    978-1-83558-292-3 (Online)

    Published Date

    2024-02-04

    Editors

    Mustafa İSTANBULLU, Cukurova University

Articles

  • Open Access | Article 2024-02-04 Doi: 10.54254/2755-2721/33/20230224

    Deep learning-based sentiment analysis for social media: A focus on multimodal and aspect-based approaches

    Commonly referred to as opinion mining, sentiment analysis harnesses the power of deep learning systems to discern human emotions and subjective sentiments towards a wide array of subjects. As such, it has become an integral tool in identifying and distinguishing sentences that harbor emotional biases or trends. By systematically examining sentiment-tinged data, researchers can unearth pivotal insights that not only reflect current perspectives but also predict future behaviors and trends. This process involves intricate computational models that analyze and interpret the emotional undertones embedded within a body of text. Whether these undertones are positive, negative, or neutral, sentiment analysis allows us to delve into the subtle nuances of human communication. This ability to "understand" and quantify sentiment is particularly vital in our modern digital age, where opinions and reviews shared through social media and online platforms can greatly influence public sentiment and consumer behavior. By extending beyond the literal meanings of words and phrases, sentiment analysis can provide a more comprehensive understanding of how people truly feel. It is instrumental in fields as diverse as marketing, politics, social science, and even artificial intelligence development, given its potential to gauge public opinion and predict societal trends. This paper aims to consolidate relevant research within the field of sentiment analysis conducted in recent years. Furthermore, it seeks to prognosticate the future trajectories and impacts of this rapidly evolving domain. Emphasis is placed on the role of deep learning and its transformational effects on the approach and capabilities of sentiment analysis, anticipating how its further advancement will continue to refine this intricate process of emotion recognition and interpretation.

  • Open Access | Article 2024-02-04 Doi: 10.54254/2755-2721/33/20230225

    Sentiment analysis in online education: An analytical approach and application

    This paper presents a groundbreaking approach to the application of sentiment analysis within the domain of online education. By introducing an innovative methodology, the aim is to streamline the process of automatically evaluating sentiments and extracting opinions from the vast sea of content produced by learners during their online interactions. This not only aids educators in swiftly gauging the general mood and perspective of their student body, but also allows them to delve deeper into the nuanced feedback provided, thus ensuring the continual improvement of course quality. In an era where digital learning platforms are growing exponentially, understanding students' attitudes, concerns, and overall satisfaction is paramount. Our methodology, therefore, is not just a technical advancement, but also a strategic tool for educational institutions aiming to thrive in the digital age. The current research landscape, while expansive, has often overlooked the significance of real-time sentiment analysis in e-learning environments. This study, therefore, bridges an important gap, bringing to the forefront the importance of harnessing student feedback in a digital format, allowing educators to tailor their approach for optimal student engagement and success.

  • Open Access | Article 2024-02-04 Doi: 10.54254/2755-2721/33/20230226

    Analyzing sentiment and its application in deep learning: Consistent behavior across multiple occasions

    This article offers a systematic review of the evolution in sentiment analysis techniques, moving from unimodal to multimodal to multi-occasion methodologies, with an emphasis on the integration and application of deep learning in sentiment analysis. Firstly, the paper presents the theoretical foundation of sentiment analysis, including the definition and classification of affect and emotion. It then delves into the pivotal technologies used in unimodal sentiment analysis, specifically within the domains of text, speech, and image analysis, examining feature extraction, representation, and classification models. Subsequently, the focus shifts to multimodal sentiment analysis. The paper offers a survey of widely utilized multimodal sentiment datasets, feature representation and fusion techniques, as well as deep learning-based multimodal sentiment analysis models such as attention networks and graph neural networks. It further addresses the application of these multimodal sentiment analysis techniques in social media, product reviews, and public opinion monitoring. Lastly, the paper underscores that challenges persist in the area of multimodal sentiment fusion, including data imbalance and disparities in feature expression. It calls for further research into cross-modal feature expression, dataset augmentation, and explainable modeling to enhance the performance of complex sentiment analysis across multiple occasions.

  • Open Access | Article 2024-02-04 Doi: 10.54254/2755-2721/33/20230227

    Vehicle detection and tracking in intelligent transportation systems based deep learning

    This research paper focuses on the advancements and optimizations made to fundamental object detection algorithms in vehicle detection. The study explores integrating and reusing CNN (Convolutional Neural Networks) models with other techniques to enhance performance. Three main models, namely Faster R-CNN (Faster Region-based Convolutional Neural Network), Improved SSD (Single Shot Multibox Detector), and YOLOv4 (You Only Look Once v4), are analyzed, showcasing their incremental improvements in accuracy and overall detection performance. However, the increased computational complexity and time demands are trade-offs. The study also presents EnsembleNet, a model combining Faster R-CNN and YOLOv5, which achieves higher average precision values. Another approach involves fusing edge features with CNN models, resulting in faster and more accurate vehicle recognition. The paper predicts future deep learning trends, emphasizing the need for improved hardware capabilities to handle complex models. Integrating deep learning with sensor fusion and edge computing holds promise for intelligent transportation systems.

  • Open Access | Article 2024-02-04 Doi: 10.54254/2755-2721/33/20230228

    Implementing the AlphaZero algorithm for Connect Four: A deep reinforcement learning approach

    The realm of board games presents a challenging domain for the application of artificial intelligence (AI), given their vast state-action space and inherent complexity. This paper explores the development of a proficient AI for Connect Four using DeepMind's AlphaZero algorithm. The algorithm employs a policy-value network for concurrent prediction of action probabilities and state values, and Monte Carlo Tree Search (MCTS) for decision-making, guided by the policy-value network. Through extensive self-play and data augmentation, our AI learns without the need for explicit prior knowledge. Our experiment demonstrated that the AI player showed significant capability in playing Connect Four, exhibiting strategic decision-making that sometimes-surpassed human performance. These results underline the potential of deep reinforcement learning in advancing AI performance in complex board games.

  • Open Access | Article 2024-02-04 Doi: 10.54254/2755-2721/33/20230229

    Deep learning applications in MRI for brain tumor detection and image segmentation

    Deep learning holds great potential in the field of MRI applications. By leveraging its advanced algorithms and neural networks, it can effectively analyze and interpret intricate patterns in medical images, aiding in precise disease detection, segmentation, and classification. Integrating deep learning techniques with MRI technology is expected to revolutionize radiology practice, facilitating enhanced diagnostic accuracy and customized treatment strategies, ultimately leading to improved patient outcomes. This article provides an overview of of the latest advancements in deep learning techniques applied to magnetic resonance imaging, specifically focusing on brain tumor detection and segmentation. The study examines eight different deep learning methods, including a multi-scale convolutional neural network, U-Net-based fully convolutional networks, cascaded anisotropic convolutional neural networks, missing modality-based tumor segmentation, Hough-CNN for deep brain region segmentation, k-Space deep learning for accelerated MRI, Multi-level Kronecker Convolutional Neural Network, and a heuristic approach for clinical brain tumor segmentation. Each method is analyzed, highlighting its specific techniques, advantages, and limitations. The comparative performance of these methods in terms of accuracy and efficiency, addressing key factors such as computational requirements, training time, and robustness, was discussed in this article. By assessing the merits and limitations of different approaches, this review seeks to offer valuable perspectives on effective utilization of deep learning techniques in clinical MRI settings for detecting and delineating brain tumors.

  • Open Access | Article 2024-02-04 Doi: 10.54254/2755-2721/33/20230230

    Data correlation and causal analysis for traffic flow prediction

    Globally, traffic congestion has become a major issue due to several issues, including the rapid urban population increase, deteriorating infrastructure, improper and disorganized traffic signal timing, and a lack of real-time data. According to INRIX, a well-known provider of traffic data and analytics, the effects of this problem on U.S. travelers in 2017 were astronomical, totaling $305 billion in wasted fuel, lost time, and increased transportation costs in congested locations. Given the limitations of building new roads, communities must investigate cutting-edge tactics and technology to ease traffic while taking practical and economical restraints into account. This study employs the Granger causality test on a dataset of 48,120 entries, primarily focusing on the variables: number of VEHICLEs and number of intersection JUNCTIONs. The objective is to ascertain the potential mutual influence between these two variables. Initial results indicate a two-way Granger-causality between the variables, implying a feedback relationship. This discovery is fundamental in understanding traffic data dynamics and could be instrumental in enhancing traffic data prediction models.

  • Open Access | Article 2024-02-04 Doi: 10.54254/2755-2721/33/20230231

    Converting graphs to ASCII art with convolutional neural network

    In an era marked by data's rapid proliferation, novel data representation and analysis methods have become increasingly significant. However, translating complex graphical structures into character-based representations remains a largely unexplored territory. This problem holds considerable importance due to its potential applications in fields like data compression and the development of innovative graphical interfaces. This study seeks to address this gap by proposing a unique methodology that uses a Convolutional Neural Network (CNN) model to translate graphical images into corresponding character arrangements. The approach involves preprocessing graphical inputs using edge detection techniques, slicing the pre-processed graph into specific columns, and feeding the resulting slices into the trained Convolutional Neural Network (CNN) model for character prediction. I interpret the SoftMax output of the model to determine the most probable character for each slice. The results indicate that the granularity of slicing impacts the accuracy of the generated character-laden graph, with higher granularity producing more precise translations. This finding demonstrates the model's ability to effectively translate graphical data into character-based representations, offering promising prospects for future study in this domain.

  • Open Access | Article 2024-02-04 Doi: 10.54254/2755-2721/33/20230235

    A review-based approach to user profiling

    With the popularity of social media, sentiment analysis and text categorisation by analysing the information people post online has become an effective method to study personality prediction. This paper focuses on how to use a personality prediction model based on Bidirectional LSTM for personality prediction. Accurate personality prediction results can provide personalised recommendation services for individuals, which has certain commercial value. In this paper, the dataset of Kaggle is first preprocessed, and then the Bidirectional LSTM model is constructed and the hyperparameters are set.The processed data are then put into the model for training and testing. Finally, the above steps are repeated using other different machine learning models. After comparison experiments with other common machine learning models, it was found that the Bidirectional LSTM model showed significant advantages in the personality prediction task, and its accuracy reached 93.5%, which was significantly higher than the traditional machine learning model.

  • Open Access | Article 2024-02-04 Doi: 10.54254/2755-2721/33/20230236

    Subway network optimization and passenger travel experience

    This conference paper examines the significance of subway network optimization in relation to passenger travel experience. It begins with an introduction that highlights the importance of subway networks in urban transportation and establishes the objectives of network optimization. A comprehensive literature review explores previous research on subway network optimization and passenger travel experience, identifying strengths, limitations, and research gaps. The paper then explores various methods of subway network optimization, including network structure optimization, train scheduling optimization, and station layout optimization, providing explanations of their principles and application scenarios. The impact of these methods on passenger travel experience is discussed, considering evaluation metrics such as crowding level, punctuality, transfer efficiency, and comfort. A case study of the China Metro system is then presented to illustrate the implementation process of network optimization and an analysis of its impact on passenger travel experience using evaluation metrics. The findings emphasize the importance of subway network optimization and its potential to enhance passenger satisfaction, connectivity, and sustainable urban development. The conclusion connects with the paper thesis and main findings that underscore the significance of subway network optimization, proposing future research directions and exploring the relationship between optimization efforts and passenger travel experience. Therefore, the paper will help understand why subway network optimization and gives insights for enhancing passenger travel experience in urban subway systems.

  • Open Access | Article 2024-02-04 Doi: 10.54254/2755-2721/33/20230237

    Optimizing traffic light signal timing using a yellow interval model: Methodology and empirical evaluation

    All over the world, especially in urban areas, the growing population and prosperity of the car industry have resulted in a steady increment in vehicles on the road, leading that already-built car-transportation infrastructure can not be commensurate with the traffic demand. Consequently, this will cause traffic congestion and thus a higher rate of traffic accidence. In such circumstances, the ability of traffic management becomes indispensable and crucial to ameliorate or even solve this problem. Although various advanced approaches can provide effective traffic management, the cost can be prohibitive to execute them pervasively. Many algorithms, especially machine learning related, require a mass of input data and a high hash rate. As a result, massive sensor systems, advanced computers, and elaborate programming will be unavoidable, which are costly. Two models in this paper, respectively, yellow light interval and traffic-light signalization, are based on summarizing and improving on previous models and are easy to understand and apply. The model focuses on timing the traffic light to maximize the number of vehicles passing through a traffic intersection in a given time. Many authors in their paper on traffic-light signalization hardly consider the effect of yellow light intervals. This paper provides a relatively comprehensive model of traffic-light signalization. Although the model in this paper is inferior to those advanced models utilized in effectiveness, this model can be applied simpler than others.

  • Open Access | Article 2024-02-04 Doi: 10.54254/2755-2721/33/20230238

    Finite element analysis and optimal design of robotic arm

    With the rapid development of modern industry and Robotics, industrial robots are widely used in manufacturing. However, because of the high price of industrial robot systems on the market, industrial robots are limited to the production of small and medium-sized enterprises. Against this background, this paper fully investigates the development process and research status of industrial manipulators at home and abroad, and designs a six-degree-of-freedom economic industrial manipulator. ANSYS Workbench software is used to analyze the statics of the arm and the whole machine, and the equivalent stress cloud map and the displacement and deformation cloud chart are obtained. The static strength and stiffness of the large arm and the whole machine are analyzed, and the reliability of the structure is verified. Then adopting the optimization module in the software, the multi-objective optimization design of the large arm is carried out. The size of the structure is reduced and the quality is reduced by 14.7% in the case that the large arm meets the requirements of the allowable stress and the maximum displacement. The target of the optimization design is realized.

  • Open Access | Article 2024-02-04 Doi: 10.54254/2755-2721/33/20230239

    The future prospects of deep learning and neural networks: Artificial intelligence's impact on education

    Artificial Intelligence (AI) has transformed a variety of areas, and education is no exception. With the development of deep learning and neural network, AI is poised to change the way people teach and learn. This paper explores the future prospects of deep learning and neural networks in education, highlighting the potential benefits and challenges they may bring. AI technologies, like deep learning algorithms and neural networks, have the potential to transform education through customized learning experiences, intelligent tutoring, streamlining administrative duties, and facilitating data-based decision making. Enhanced personalized learning helps students to learn at their own pace and in their preferred style, smart tutoring systems offer personalized guidance and support. Automation of administrative tasks increases efficiency and accuracy, while data-driven decision making helps educators make informed choices about students' outcomes. However, the implementation of AI in education poses challenges such as data privacy, equity, and the preservation of the teacher-student relationship. Efforts should be made to address these challenges and fully harness the potential of deep learning and neural networks in education.

  • Open Access | Article 2024-02-04 Doi: 10.54254/2755-2721/33/20230240

    Application of image recognition technology based on deep learning in plants disease detection and diagnosis

    With increasing demographic trends, the world will likely confront a looming food crisis in the coming decades. This paper points out the world's food crisis brought by the increasing population and the agricultural sector's challenges. Among them, the increase in plant diseases is considered one of the leading causes of agricultural inefficiency and food production problems. For example, in many countries, cash crops such as jute, pomegranate, and tomato have been severely affected by plant diseases, resulting in reduced yields and, thus, increased economic losses. Traditional disease detection methods, such as manual visual inspection and laboratory analysis, are inaccurate and time-consuming, requiring a lot of manpower and capital investment. In recent years, researchers have begun to apply deep learning-based image recognition technology to detect and diagnose plant diseases to improve accuracy and efficiency. This paper will lay the foundation for further research by summarizing the existing research and analyzing the advantages and applicability of different application methods.

  • Open Access | Article 2024-02-04 Doi: 10.54254/2755-2721/33/20230242

    A review on current progress of semantic segmentation

    Semantic segmentation, as an important task in the field of computer vision, has wide applications in image analysis and scene analysis. These application domains include autonomous driving, medical image analysis, image identification, and intelligent video surveillance. However, it faces many challenges due to the complex image structures and some confusing relationships between objects. This paper aims to provide an overview of key concepts in the field of semantic segmentation, including datasets and annotations, data augmentation, some relevant algorithms and models, and loss functions. By introducing and analyzing these concepts, we can gain a comprehensive understanding of the research progress and future directions in semantic segmentation. This paper also provides research advancements in the field of semantic segmentation. Through the introduction and analysis of these different concepts, we gain a deeper understanding of the current state and challenges of semantic segmentation. With the continuous development of deep learning techniques, we can expect semantic segmentation to have broader applications in the fields of computer vision and artificial intelligence.

  • Open Access | Article 2024-02-04 Doi: 10.54254/2755-2721/33/20230245

    Accurate and efficient galaxy classification based on mobile vision transformer

    Understanding the formation and evolution of galaxies in observational cosmology heavily relies on galaxy morphological classification. Nevertheless, the continuously growing volume of astronomical data has surpassed human capacity for manual classification. In this context, deep learning presents a promising approach to enhancing classifying galaxies. In this paper, the Mobile Vision Transformer (MobileViT) is introduced to construct an efficient and accurate galaxy classifier. Transfer learning is introduced to assist in model fine-tuning. MobileViT combines the features of MobileNet and Visual Transformer (ViT). A lightweight model is used to effectively analyse the relationships between sequences for efficient and accurate classification. Experiments are built on Galaxy10 DECals dataset. Excellent performance is achieved in identifying galaxy types compared to other lightweight models. The model achieves an accuracy of over 87% and maintains a high speed of inference of less than 50 milliseconds per step. Experimental results show that the introduction of MobileViT is the best solution for efficient galaxy classification. The model can be deployed on any portable device for instant observation and classification.

  • Open Access | Article 2024-02-04 Doi: 10.54254/2755-2721/33/20230246

    Harnessing the power of AutoML: A comparative study of image recognition techniques for smoking detection

    As infrastructure continues to evolve, the significance of fire protection escalates. Many fires are caused by smoking in smoke-free areas, underscoring the necessity to promptly detect smoking activities in hazardous zones. In this scenario, image recognition emerges as a pivotal tool. The accuracy and efficiency of image recognition bear substantial implications for both academic and industrial sectors, and these aspects form the crux of our investigation. This study aims to compare the performance of image recognition techniques based on automatic machine learning with those of traditional methods such as YOLO. Our findings indicate that image recognition powered by automatic machine learning outperforms YOLO recognition in terms of efficiency and accuracy.

  • Open Access | Article 2024-02-04 Doi: 10.54254/2755-2721/33/20230248

    Bio-inspired soft robot with varied localized stiffness

    Soft robots are a type of intelligent robot with high adaptability, and the majority of them are made from soft materials, so they are flexible and adaptable. The variable stiffness function of soft robots is crucial, as it can enhance the robot's adaptability, safety, and dependability. By adjusting the stiffness, the soft robot is able to maintain a stable motion state in a complex environment, reduce environmental interference, and perform human-like actions with improved target control. The elephant trunk served as inspiration for the way proposed in this study for modifying the soft robot's arm's stiffness. By changing the insertion scheme of the interference plate in a flexible manner, the robot arm can have variable bending capability, allowing it to complete a variety of work instructions given by humans and to satisfy a variety of work requirements.

  • Open Access | Article 2024-02-04 Doi: 10.54254/2755-2721/33/20230255

    Research and application analysis of typical recommendation methods

    Recommendation systems are vital tools in the modern digital landscape for handling vast amounts of online data and are now employed across a variety of sectors. This manuscript underlines the importance of video recommendation algorithms and delves into the three main types: content-based, collaborative filtering, and hybrid algorithms. Content-based algorithms offer suggestions by assessing and matching the descriptive attributes of videos. These attributes could range from the genre, the actors involved, the director, or any other related metadata. On the other hand, collaborative filtering algorithms provide recommendations by comparing the user's historical data with that of others. They function on the principle that individuals who have agreed in the past are likely to agree again in the future. Hybrid recommendation algorithms, as the name suggests, are a combination of both content-based and collaborative filtering algorithms. They aim to harness the strengths of the two while mitigating their individual weaknesses, providing more balanced and accurate recommendations. This detailed exploration of these algorithms not only underscores their significance but also paves the way for the future, heralding the potential advancements and adaptations these systems could undergo. Continued innovation in this realm has the potential to revolutionize how information is processed and shared, ultimately enriching the user experience across the digital platform landscape.

  • Open Access | Article 2024-02-04 Doi: 10.54254/2755-2721/33/20230256

    Research of artificial intelligence in imperfect information card games

    Artificial intelligence (AI) in games has advanced significantly, notably in perfect information games such as Go and Chess. Imperfect information games, in which participants do not have complete information about the game state, create more difficulties. They incorporate both public and private observations, where strategies must be improved to achieve a Nash equilibrium. This study investigates artificial intelligence and reinforcement learning approaches, in which agents learn to maximize future rewards through interactions with their surroundings. The paper then focuses on card game research platforms such as RLCard and OpenAI Gym. It gives a comprehensive summary of research in No Limit Texas Hold'em, a difficult two-player poker game with a large decision space. DeepStack and Libratus are successful systems that have attained expert-level and superhuman play, respectively. Pluribus, a superhuman artificial intelligence for six-player poker, and DouZero, a pure reinforcement learning technique for the multiplayer card game, DouDiZhu, are both investigated. Overall, this paper provides background information on reinforcement learning and imperfect information games, analyzes commonly used research platforms, evaluates the effectiveness of AI algorithms in various card games, and offers future research areas and directions.

Copyright © 2023 EWA Publishing. Unless Otherwise Stated