• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.23 no.6
• /
• pp.61-67
• /
• 2023

The rise of 5G and the proliferation of smart devices have underscored the significance of multi-access edge computing (MEC). Amidst this trend, interest in effectively processing computation-intensive and latency-sensitive applications has increased. This study investigated a novel task offloading strategy considering the probabilistic MEC environment to address these challenges. Initially, we considered the frequency of dynamic task requests and the unstable conditions of wireless channels to propose a method for minimizing vehicle power consumption and latency. Subsequently, our research delved into a deep reinforcement learning (DRL) based offloading technique, offering a way to achieve equilibrium between local computation and offloading transmission power. We analyzed the power consumption and queuing latency of vehicles using the deep deterministic policy gradient (DDPG) and deep Q-network (DQN) techniques. Finally, we derived and validated the optimal performance enhancement strategy in a vehicle based MEC environment.

• Advanced Industrial SCIence
• /
• v.3 no.2
• /
• pp.17-22
• /
• 2024

Recently, as the ICT field has been used in various environments, it has become possible to analyze pests by crops, use robots when harvesting crops, and predict by big data by utilizing ICT technologies in a sustainable agricultural environment. However, in a sustainable agricultural environment, efforts to solve resource depletion, agricultural population decline, poverty increase, and environmental destruction are constantly being demanded. This paper proposes an artificial intelligence-based big data processing analysis method to reduce the production cost and increase the efficiency of crops based on a sustainable agricultural environment. The proposed technique strengthens the security and reliability of data by processing big data of crops combined with AI, and enables better decision-making and business value extraction. It can lead to innovative changes in various industries and fields and promote the development of data-oriented business models. During the experiment, the proposed technique gave an accurate answer to only a small amount of data, and at a farm site where it is difficult to tag the correct answer one by one, the performance similar to that of learning with a large amount of correct answer data (with an error rate within 0.05) was found.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.24 no.1
• /
• pp.1-8
• /
• 2024

On-device AI technology, which can operate AI models at the edge devices to support real-time processing and privacy enhancement, is attracting attention. As intelligent IoT is applied to various industries, services utilizing the on-device AI technology are increasing significantly. However, general deep learning models require a lot of computational resources for inference and learning. Therefore, various lightweighting methods such as quantization and pruning have been suggested to operate deep learning models in embedded edge devices. Among the lightweighting methods, we analyze how to lightweight and apply deep learning models to edge computing devices, focusing on pruning technology in this paper. In particular, we utilize dynamic and static pruning techniques to evaluate the inference speed, accuracy, and memory usage of a lightweight AI vision model. The content analyzed in this paper can be used for intelligent video control systems or video security systems in autonomous vehicles, where real-time processing are highly required. In addition, it is expected that the content can be used more effectively in various IoT services and industries.

• Korean Journal of Radiology
• /
• v.24 no.11
• /
• pp.1061-1080
• /
• 2023

Artificial intelligence (AI) in radiology is a rapidly developing field with several prospective clinical studies demonstrating its benefits in clinical practice. In 2022, the Korean Society of Radiology held a forum to discuss the challenges and drawbacks in AI development and implementation. Various barriers hinder the successful application and widespread adoption of AI in radiology, such as limited annotated data, data privacy and security, data heterogeneity, imbalanced data, model interpretability, overfitting, and integration with clinical workflows. In this review, some of the various possible solutions to these challenges are presented and discussed; these include training with longitudinal and multimodal datasets, dense training with multitask learning and multimodal learning, self-supervised contrastive learning, various image modifications and syntheses using generative models, explainable AI, causal learning, federated learning with large data models, and digital twins.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.24 no.2
• /
• pp.177-184
• /
• 2024

In the situation of increasing data at fast speed, we use all kinds of complex ensemble and deep learning algorithms to get the highest accuracy. It's sometimes questionable how these models predict, classify, recognize, and track unknown data. Accomplishing this technique and more has been and would be the goal of intensive research and development in the data science community. A variety of reasons, such as lack of data, imbalanced data, biased data can impact the decision rendered by the learning models. Many models are gaining traction for such interpretations. Now, LIME and SHAP are commonly used, in which are two state of the art open source explainable techniques. However, their outputs represent some different results. In this context, this study introduces a coupling technique of LIME and Shap, and demonstrates analysis possibilities on the decisions made by LightGBM and Keras models in classifying a transaction for fraudulence on the IEEE CIS dataset.

• Journal of Microbiology and Biotechnology
• /
• v.34 no.5
• /
• pp.1017-1028
• /
• 2024

Lignocellulolytic enzymes play a crucial role in efficiently converting lignocellulose into valuable platform molecules in various industries. However, they are limited by their production yields, costs, and stability. Consequently, their production by producers adapted to local environments and the choice of low-cost raw materials can address these limitations. Due to the large amounts of olive stones (OS) generated in Morocco which are still undervalued, Penicillium crustosum, Fusarium nygamai, Trichoderma capillare, and Aspergillus calidoustus, are cultivated under different fermentation techniques using this by-product as a local lignocellulosic substrate. Based on a multilevel factorial design, their potential to produce lignocellulolytic enzymes during 15 days of dark incubation was evaluated. The results revealed that P. crustosum expressed a maximum total cellulase activity of 10.9 IU/ml under sequential fermentation (SF) and 3.6 IU/ml of β-glucosidase activity under submerged fermentation (SmF). F. nygamai recorded the best laccase activity of 9 IU/ml under solid-state fermentation (SSF). Unlike T. capillare, SF was the inducive culture for the former activity with 7.6 IU/ml. A. calidoustus produced, respectively, 1,009 ㎍/ml of proteins and 11.5 IU/ml of endoglucanase activity as the best results achieved. Optimum cellulase production took place after the 5th day under SF, while ligninases occurred between the 9th and the 11th days under SSF. This study reports for the first time the lignocellulolytic activities of F. nygamai and A. calidoustus. Furthermore, it underlines the potential of the four fungi as biomass decomposers for environmentally-friendly applications, emphasizing the efficiency of OS as an inducing substrate for enzyme production.

• The Journal of the Institute of Internet, Broadcasting and Communication
• /
• v.24 no.4
• /
• pp.57-64
• /
• 2024

The introduction of Smart Factories has transformed manufacturing towards more objective and efficient line management. However, most companies are not effectively utilizing the vast amount of sensor data collected every second. This study aims to use this data to predict product quality and manage production processes efficiently. Due to security issues, specific sensor data could not be verified, so semiconductor process-related training data from the "SAMSUNG SDS Brightics AI" site was used. Data preprocessing, including removing missing values, outliers, scaling, and feature elimination, was crucial for optimal sensor data. Oversampling was used to balance the imbalanced training dataset. The SVM (rbf) model achieved high performance (Accuracy: 97.07%, GM: 96.61%), surpassing the MLP model implemented by "SAMSUNG SDS Brightics AI". This research can be applied to various topics, such as predicting component lifecycles and process conditions.

• KSII Transactions on Internet and Information Systems (TIIS)
• /
• v.18 no.9
• /
• pp.2692-2716
• /
• 2024

Over the past decade, law enforcement organizations have been dealing with the development of cybercrime. To address this growing problem, law enforcement organizations apply various digital forensic (DF) tools and techniques to investigate crimes involving digital devices. This ensures that evidence is admissible in legal proceedings. Consequently, DF analysts may need to invest more in proprietary DF hardware and software to maintain the viability of the DF lab, which will burden budget-constrained organizations. As an alternative, the open source DF tool is considered a cost-saving option. However, the admissibility of digital evidence obtained from these tools has yet to be tested in courts, especially in Malaysia. Therefore, this study aimed to explore the admissibility of digital evidence obtained through open source DF tools. By reviewing the existing literature, the factors that affect the admissibility of the evidence produced by these tools in courts were identified. Further, based on the findings, a conceptual framework was developed to ensure the admissibility of the evidence so that it will be accepted in the court of law. This conceptual framework was formed to outline the factors affecting the admissibility of digital evidence from open source DF tools, which include; 1) The Availability and Capability of open source DF tools, 2) the Reliability and Integrity of the digital evidence obtained from open source DF tools, 3) the Transparency of the open source DF tools, and 4) the Lack of Reference and Standard of open source DF tools. This study provides valuable insights into the digital forensic field, and the conceptual framework can be used to integrate open source DF tools into digital forensic investigations.

• International conference on construction engineering and project management
• /
• 2024.07a
• /
• pp.455-462
• /
• 2024

In the construction industry, known for its dynamic and hazardous environments, there exists a crucial demand for effective safety incident prevention. Traditional approaches to monitoring on-site safety, despite their importance, suffer from being laborious and heavily reliant on subjective, paper-based reports, which results in inefficiencies and fragmented data. Additionally, the incorporation of computer vision technologies for automated safety monitoring encounters a significant obstacle due to the lack of suitable training datasets. This challenge is due to the rare availability of safety accident images or videos and concerns over security and privacy violations. Consequently, this paper explores an innovative method to address the shortage of safety-related datasets in the construction sector by employing generative artificial intelligence (AI), specifically focusing on the Stable Diffusion model. Utilizing real-world construction accident scenarios, this method aims to generate photorealistic images to enrich training datasets for safety surveillance applications using computer vision. By systematically generating accident prompts, employing static prompts in empirical experiments, and compiling datasets with Stable Diffusion, this research bypasses the constraints of conventional data collection techniques in construction safety. The diversity and realism of the produced images hold considerable promise for tasks such as object detection and action recognition, thus improving safety measures. This study proposes future avenues for broadening scenario coverage, refining the prompt generation process, and merging artificial datasets with machine learning models for superior safety monitoring.

• The Journal of the Korea institute of electronic communication sciences
• /
• v.19 no.5
• /
• pp.941-950
• /
• 2024

This study employed network analysis techniques to examine the relationships between spatiotemporal factors associated with crimes in public restrooms, drawing on diverse relevant data sources. We then evaluated the relative importance of these factors in crime occurrence. Variables related to crime incidence were identified, and their interconnectedness was assessed for network analysis, resulting in a data-driven network model with complex relational structures. The network model contributed to calculating the weight of each factor and identifying key elements. The location of public restrooms, usage time, surrounding environment, and facility conditions emerged as crucial factors in crime occurrence, with lighting quality and local security status showing high weightings. These findings can be utilized to prioritize interventions in public restroom design and management to enhance safety. The network analysis methodology demonstrated its potential in proposing crime prevention measures for public spaces, including restrooms, and contributing to the creation of safer public environments.