Publication Search

This study aims to develop and implement an artificial intelligence-based reading learning application using Deep Learning technology to enhance the literacy skills of eighth-grade junior high school students. The research employed the Kemmis & McTaggart Classroom Action Research model combined with a mixed-methods approach. Data collection involved pretests and posttests, complemented by observations, interviews, and questionnaires. The findings revealed that the use of this application significantly improved students' reading comprehension, question-answering skills, and overall engagement in the learning process. Key features of the application, such as adaptive learning technology, allowed for real-time adjustments to the difficulty level of the material, which catered to each student’s individual learning pace. Additionally, the provision of instant feedback enhanced the learning experience by helping students understand their progress and areas for improvement. These results suggest that the application is an effective tool in fostering literacy development and aligns with the goals of the Independent Curriculum. Consequently, this Deep Learning-based application offers a promising innovation for improving student literacy skills in the digital age. 

Object detection plays a crucial role in intelligent transportation systems, particularly for outdoor traffic monitoring applications that require accurate and real-time performance under limited computational resources. Recent developments in YOLO-based architectures have introduced multiple model variants; however, their practical performance under constrained training conditions remains insufficiently explored. This study presents a comparative evaluation of YOLOv5, YOLOv7, and YOLOv8 for outdoor traffic object detection using a real-world dataset and identical experimental settings. The main objective of this research is to analyze the robustness and detection quality of different YOLO variants when trained with a limited number of epochs, reflecting practical deployment scenarios. All models were trained and evaluated using the same dataset, preprocessing pipeline, and hardware configuration to ensure a fair comparison. Performance evaluation was conducted using multiple metrics, including precision, recall, mAP@50, Precision–Recall curves, area under the curve (AUC), and peak F1-score. Experimental results indicate that YOLOv5 outperformed YOLOv7 and YOLOv8 in terms of overall detection stability and robustness. The merged Precision–Recall analysis shows that YOLOv5 achieved a higher effective AUC and superior mAP@50, reflecting better global detection performance. In addition, YOLOv5 exhibited a higher peak F1-score, indicating a more balanced trade-off between precision and recall. In contrast, YOLOv7 and YOLOv8 showed performance degradation under limited training conditions despite their more advanced architectures. These findings suggest that YOLOv5 remains a reliable and efficient solution for outdoor traffic object detection, particularly in resource-constrained environments. The study highlights the importance of comprehensive evaluation metrics and practical experimental settings when selecting object detection models for real-world applications.

Object detection plays a crucial role in intelligent transportation systems, particularly for outdoor traffic monitoring applications that require accurate and real-time performance under limited computational resources. Recent developments in YOLO-based architectures have introduced multiple model variants; however, their practical performance under constrained training conditions remains insufficiently explored. This study presents a comparative evaluation of YOLOv5, YOLOv7, and YOLOv8 for outdoor traffic object detection using a real-world dataset and identical experimental settings. The main objective of this research is to analyze the robustness and detection quality of different YOLO variants when trained with a limited number of epochs, reflecting practical deployment scenarios. All models were trained and evaluated using the same dataset, preprocessing pipeline, and hardware configuration to ensure a fair comparison. Performance evaluation was conducted using multiple metrics, including precision, recall, mAP@50, Precision–Recall curves, area under the curve (AUC), and peak F1-score. Experimental results indicate that YOLOv5 outperformed YOLOv7 and YOLOv8 in terms of overall detection stability and robustness. The merged Precision–Recall analysis shows that YOLOv5 achieved a higher effective AUC and superior mAP@50, reflecting better global detection performance. In addition, YOLOv5 exhibited a higher peak F1-score, indicating a more balanced trade-off between precision and recall. In contrast, YOLOv7 and YOLOv8 showed performance degradation under limited training conditions despite their more advanced architectures. These findings suggest that YOLOv5 remains a reliable and efficient solution for outdoor traffic object detection, particularly in resource-constrained environments. The study highlights the importance of comprehensive evaluation metrics and practical experimental settings when selecting object detection models for real-world applications.

Teachers’ managerial competence plays a crucial role in ensuring effective and relevant learning, particularly in vocational education. Teachers are required not only to master subject matter but also to plan, manage, and evaluate learning through well-structured and contextual learning tools. This Community Service activity aimed to strengthen teachers’ managerial competence through training on the development of learning tools using a deep learning approach at SMK Muhammadiyah 2 Tangerang Selatan. The implementation method included conceptual material delivery, practical training, and mentoring in designing learning tools. The deep learning approach was applied to encourage learning designs that emphasize deep understanding, critical and reflective thinking, and alignment with workplace demands. The results showed an improvement in teachers’ understanding and skills in developing more systematic, innovative, and adaptive learning tools aligned with 21st-century competencies. This activity contributed positively to strengthening teachers’ roles as learning managers and supporting the improvement of learning quality in vocational schools.

The increasing sedentary lifestyle in the digital era has the potential to cause various health problems due to lack of physical activity. One approach that can be taken to encourage physical activity is through the use of digital games with body movement-based control mechanisms. This study aims to develop a body gesture-based game character control system using a hybrid Convolutional Neural Network (CNN) and Long Short-Term Memory (LSTM) model. CNN is used to extract spatial features from each video frame, while LSTM serves to model the temporal relationship between frames so that movement patterns can be recognized sequentially. The research method used refers to the Machine Learning Lifecycle stages, starting from data collection, preprocessing, model development, to implementation in the endless runner game genre. Testing results show that the CNN–LSTM model is capable of classifying body gestures and generating outputs that can be used as commands to control game characters. The implementation of this system enables more natural and interactive game interactions without conventional input devices, and has the potential to encourage players to lead a more active lifestyle.

This study aims to develop an Indonesian traffic sign detection system using a transfer learning approach to improve road safety and traffic efficiency. The dataset was obtained from Kaggle and consists of 2,100 images across 21 traffic sign classes. The research stages include data collection, preprocessing to reduce noise and normalize image brightness, object detection using YOLOv5, and classification based on transfer learning with ResNet, VGG-16, and MobileNet architectures. Model performance was evaluated using accuracy, precision, recall, and F1-score metrics. Experimental results indicate that the YOLOv5 model is capable of detecting traffic sign objects; however, the classification performance remains relatively low, with a mean Average Precision (mAP) value of 0.17. These findings suggest that further optimization is required in data preprocessing, dataset quality, and model parameter tuning to achieve better performance. This study demonstrates that transfer learning has significant potential for developing computer vision-based traffic sign detection systems, although further improvements are necessary to ensure robustness under real-world Indonesian traffic conditions.

This study is motivated by the increasing need for accurate modeling and classification of one-dimensional signal data in intelligent systems. The rapid development of deep learning has led to the adoption of more adaptive and complex neural network architectures capable of capturing both temporal dependencies and local patterns in sequential data. This research aims to analyze and compare the performance of several deep learning models, namely Recurrent Neural Network (RNN), Long Short-Term Memory (LSTM), Gated Recurrent Unit (GRU), and a hybrid Convolutional Neural Network–GRU (CNN–GRU) model for signal data classification. The research method employs a quantitative experimental approach involving data preprocessing, windowing, model training, and performance evaluation. Model performance is evaluated using accuracy, precision, recall, and F1-score metrics. The results indicate that the hybrid CNN–GRU model outperforms the other models, particularly in capturing local features and long-term temporal dependencies within signal data. These findings suggest that the integration of convolutional layers and recurrent mechanisms enhances feature representation and learning stability. This study is expected to contribute both theoretically and practically to the development of deep learning models for signal processing and time-series-based intelligent applications.

This research aims to develop a deep learning-based mathematics e-module with the integration of the context of the Suramadu Bridge in scale and comparison materials for grade VII junior high school students. The development model used is the 4D (Define, Design, Develop, Disseminate) model which includes the stage of defining learning needs, designing e-modules, product development, and limited deployment. The research instruments used included validation sheets of media and material experts, observation sheets of teacher and student activities, learning outcome tests, and student response questionnaires. The results of the study showed that the e-modules developed met the valid criteria with a media validity level of 94.29% and material of 95%. In addition, the e-module is considered practical with teacher practicality of 78.67% and students of 86.67%, and effective with a learning effectiveness rate of 83.83%. The students' response to the e-module was also very positive, which showed that the integration of the local context of the Suramadu Bridge and the deep learning approach was able to increase student engagement, learning motivation, and understanding of mathematical concepts in a meaningful way. These findings indicate that local context-based e-modules can be an innovative alternative in mathematics learning that is relevant to students' real lives as well as support the implementation of 21st century learning.

Islamic education in the digital era faces significant challenges in improving the quality of learning and developing students' critical thinking skills. One promising solution is the implementation of deep learning-based Islamic education, which can facilitate the teaching of religious material more effectively and easily accessible via the internet. Deep learning has the potential to strengthen students' critical thinking skills and understanding of Islamic values ​​in a more interactive and personalized manner. This study used a literature review method to analyze various relevant literature sources. The results indicate that deep learning can increase student engagement, enable personalized learning, and provide real-time evaluation of learning progress. However, the implementation of this technology also faces challenges, particularly related to the availability and quality of relevant data. Therefore, efforts are needed to improve data quality and develop learning models tailored to student needs to achieve optimal results in technology-based Islamic education.

The increasing demand for low latency and high-throughput multimedia applications has spurred significant advancements in hardware software co design. This study explores the integration of custom digital signal processing (DSP) hardware accelerators with optimized software frameworks to enhance deep learning accelerated DSP tasks. The proposed co design approach significantly reduces latency and improves throughput compared to traditional software-only DSP implementations. Through the development of custom hardware accelerators built with FPGA technology, the system achieves up to a 1.85x reduction in latency and a 1.5x improvement in throughput for real-time multimedia tasks such as image recognition, video decoding, and audio processing. The combination of hardware and software optimizations allows for better resource utilization, enabling the parallel processing of computationally intensive tasks while the software framework handles less demanding operations. Additionally, the co design system demonstrated improved energy efficiency, making it highly suitable for embedded systems. The results show that the hardware software co design approach offers substantial advantages in performance, latency reduction, and energy efficiency, positioning it as a viable solution for real-time multimedia applications. The findings have important implications for applications requiring fast data processing, such as autonomous driving, healthcare, and disaster management. Future research could explore alternative hardware accelerators, advanced software optimizations, and AI-based resource management to further improve the system’s efficiency and scalability for more complex multimedia tasks.

User reviews on the Shopee e-commerce platform represent an important source of information for understanding consumer perceptions of products and services. Sentiment analysis is commonly applied to classify user opinions into positive, neutral, and negative sentiment categories based on textual data. This study aims to analyze the performance of the Long Short-Term Memory (LSTM) method in sentiment classification of Shopee user reviews. The dataset used in this study consists of Indonesian-language user reviews that have undergone preprocessing stages, including case folding, text cleaning, tokenization, and stopword removal. The LSTM model was trained using preprocessed text represented as word sequences. Model performance was evaluated using overall accuracy and class-wise classification results. The experimental results indicate that the LSTM method achieved an overall accuracy of 87.62%. In addition, the classification performance for the positive sentiment class reached 95.27%, the neutral class achieved 4.96%, and the negative class reached 74.26%. These results demonstrate that the LSTM method performs well in classifying sentiment in Shopee user reviews, particularly for positive sentiment. This study is expected to provide insights and references for the application of deep learning methods in sentiment analysis of Indonesian e-commerce review data.

This research aims to develop a multiclass flower image classification system using the Convolutional Neural Network (CNN) algorithm with the EfficientNet architecture. The main problem addressed is the difficulty of manual identification of flower species that share high visual similarity. The research stages include collecting 17,299 flower images across 19 classes, performing data preprocessing such as image resizing, pixel normalization, and augmentation, followed by model training using the EfficientNet transfer learning approach. The model was trained for 10 epochs with an 80:20 training-validation data split. The evaluation results show that the model achieved a validation accuracy of 98.05% with a loss value of 0.0968, and an average precision, recall, and F1-score of 0.98. The trained model was then implemented into a web-based application built using the Next.js framework, enabling users to upload flower images and obtain real-time classification results via the Hugging Face API. The system successfully identified flower species with a confidence level of 99.87%. These findings demonstrate that combining a modern CNN architecture with transfer learning provides efficient and highly accurate flower classification performance, which can be effectively implemented for educational and digital conservation purposes.

The development of digital technology and the demands of 21st-century learning require teachers to implement meaningful and student-centered learning through an immersive learning approach. However, many elementary school teachers still face limitations in understanding the concept of immersive learning and utilizing technology-based learning media, especially three-dimensional (3D) and Artificial Intelligence (AI)-based media. Therefore, this Field Study activity aims to improve the knowledge and skills of teachers and students in implementing immersive learning through the use of 3D-based learning media and AI technology at the UPTD of SD Negeri 155684 Lubuk Tukko 1, Pandan District. The activity was implemented through a four-day workshop involving teachers and students, using an approach of socialization, demonstrations, hands-on practice, discussion, and reflection. The workshop materials covered the concept of immersive learning, the use of 3D-based learning media, gamification of learning through 3D Media applications, and the use of Artificial Intelligence in developing teaching materials and learning media. Evaluation of the activity was conducted through pre- and post-tests, participant observation, and analysis of the resulting learning products. .The results of the activity showed an increase in teachers' understanding and skills in designing and implementing technology-based immersive learning. Teachers were able to produce interactive learning media, teaching modules, and evaluation questions using 3D media, 3D media, and AI. Furthermore, this activity also increased teachers' motivation, creativity, and awareness of the importance of digital literacy in learning. This Field Study activity made a positive contribution to improving teacher competency and supporting the creation of more innovative, interactive, and relevant learning that reflects the characteristics of 21st-century learners.

This research aims to analyze the adaptive role of a dance teacher in teaching the subject of music art at SMK Asta Mitra Purwodadi, as well as identifying the patterns of collaborative approaches and the implementation of the Deep Learning strategy in the learning process. This phenomenon attracts researchers due to the case of a teacher teaching cross-disciplinary subjects, although still within the domain of the arts. This study employs a qualitative method using observation, semi-structured interviews, and documentation techniques. The main subject of this research is Bu Evi, a dance teacher who also serves as the music art instructor. The research results indicate that the non-specialist teacher (with a dance background) successfully overcomes limitations in teaching music by performing a role transformation into a facilitator focusing on expression, rhythm, and performance. The implementation of Deep Learning is realized effectively through a kinesthetic approach, where abstract music concepts (such as tempo and dynamics) are translated through body movement exploration, which is proven to increase student enthusiasm and understanding. Furthermore, the teacher implements a strategic collaborative pattern with the music extracurricular teacher, where responsibilities are divided between aspects of expression/movement and the deepening of theory/vocal technique, ensuring talented students still receive adequate technical development. Assessment focuses on performative aspects (expression, tempo, teamwork, and self-confidence). This research concludes that the teacher's professionalism and pedagogical creativity are the main keys to bridging competency gaps, while simultaneously ensuring the achievement of the goals of the Merdeka Curriculum in shaping student character and 21st-century skills.

This research aims to determine the extent to which interactive quizzes can help improve students' understanding of Economics lessons in the Merdeka Curriculum, which focuses on a Deep Learning approach implemented at SMA Negeri 4 Surakarta. The results show that quite a few students have difficulty grasping economic concepts. The lack of participation and learning motivation, combined with the dominance of the lecture method, makes the learning process tend to be one-way. Conversely, the application of digital technology by educators is still very limited. This research uses a descriptive qualitative approach, collecting data thru observation, interviews, and documentation. A total of five students from the eleventh grade were selected based on purposive sampling techniques who had already taken economics lessons. The results show that the quiz format contributes to increasing students' interest, attention, and participation due to the presence of visual elements, screen animation, and a game-like atmosphere. However, obstacles such as slow internet connectivity or a slightly longer completion time than estimated did arise, although they did not hinder overall learning activities. Essentially, implementing this type of quiz aligns with the direction of the Merdeka Curriculum, which emphasizes in-depth conceptual understanding while also encouraging self-study habits, training the ability to reflect on the material, and creating a classroom atmosphere that tends to be more dynamic.

This study aims to describe the application of deep learning based tolerance values with a pedagogical approach in supporting classroom management in grade VI of SDN Beji 01 East Ungaran. The deep learning approach is understood as a pedagogical strategy that emphasizes the cognitive, affective, and social engagement of students through meaningful learning experiences. This study uses a descriptive qualitative approach with data collection techniques in the form of semi-structured interviews, observations, and documentation. The research subjects include grade VI teachers, principals, and students with diverse social and cultural backgrounds. The results of the study show that the application of tolerance values through deep learning strategies is able to create an inclusive, safe, and conducive classroom climate. The implementation is reflected through the practice of random seating arrangements, a culture of sharing learning media, deliberation in conflict resolution, fair study timing, and strengthening mutual respect between students. These findings confirm that the integration of deep learning-based tolerance values contributes significantly to the effectiveness of classroom management and the strengthening of the social character of elementary school students.

Reading literacy among 6th grade elementary students is often superficial, limiting their ability to analyze implicit meanings and connect texts to real-world contexts. This review examines the potential of deep learning as a pedagogical approach to enhance in-depth text comprehension. Deep learning emphasizes active engagement, reflection, and the construction of meaningful knowledge, aiming to foster critical thinking and improve comprehension. Recent studies highlight implementation strategies such as reflective journaling and interactive discussions, which have demonstrated significant improvements in students' critical thinking and comprehension scores (p < 0.05). These findings suggest that deep learning methods surpass traditional approaches by promoting higher-order cognitive skills, enabling students to analyze and interpret texts more effectively. However, challenges such as inadequate teacher training persist, which may hinder the full integration of deep learning techniques. To address these challenges, further research is needed to explore scalable digital tools that can support deep learning in diverse classroom settings. By examining the potential for digital integration, future studies could provide insights into how technology can facilitate the widespread adoption of deep learning strategies, making them more accessible and effective for a broader range of students. Ultimately, this review underscores the promise of deep learning in enhancing reading literacy and suggests that addressing the barriers to its implementation could have significant educational benefits.

Flower disease detection is a common challenge in modern agriculture. Various factors, such as changes in leaf color, shape, petal structure, and environmental conditions, make it difficult to achieve high accuracy with conventional models. Transfer learning is an effective solution to improve model performance in image detection, especially when available data is limited. This study used several pre-trained models, namely VGG16, ResNet50, and EfficientNet-B0, to detect three types of flower diseases: black spot on roses, white powdery mildew, and leaf rust. The process included data processing, increasing the data volume, model training, and result verification. The results showed that the EfficientNet-B0 model provided the highest accuracy of 97.2%, significantly better than the CNN model created from scratch with an accuracy of 85.1%. This study proves that the transfer learning method is very effective in improving the accuracy of flower disease detection. These results confirm that transfer learning is effective for detecting plant diseases with higher accuracy, especially when the dataset is limited.  

Sudden Infant Death Syndrome (SIDS) is a sudden and unexpected death in infants that is often associated with the prone sleeping position. This study aims to develop an automated monitoring system capable of detecting SIDS risk factors using the YOLOv8 algorithm and to analyze the effect of data augmentation on model performance. The dataset consists of two classes, baby-lying-on-back (supine) and baby-lying-on-stomach (prone), which were processed through model training and evaluation using precision, recall, F1-score, and mAP metrics. The model was trained under two scenarios, without data augmentation and with data augmentation. The results show that the model without augmentation achieved a precision of 90%, recall of 85%, F1-score of 86%, and mAP50 of 93.7%. After applying augmentation, performance improved to a precision of 90%, recall of 87%, F1-score of 88%, and mAP50 of 95.1%. These findings indicate that augmentation increases detection accuracy and enhances model generalization, including robustness against variations in lighting and camera angles. Furthermore, testing with image and video inputs revealed that the non-augmented model exhibited a tendency toward overfitting, particularly in favor of the baby-lying-on-stomach, whereas the augmented model successfully classified both classes accurately. The developed system is also equipped with an alarm feature and early-warning notifications via Telegram to smartphone when a prone position is detected for a certain duration. Overall, the results demonstrate that YOLOv8 with data augmentation is effective for an automated, non-invasive monitoring system for infants, making it suitable for detecting and preventing potential SIDS risk factors.

Flower disease detection is a significant challenge in modern agriculture, particularly with factors such as changes in leaf color, petal shape and structure, and environmental conditions affecting the accuracy of conventional models. These factors make it difficult to achieve optimal results using traditional methods. Transfer learning is an effective solution to improve image detection performance, especially when data is limited. This study used several pre-trained models, namely VGG16, ResNet50, and EfficientNet-B0, to detect three types of flower diseases: black spot on roses, white powdery mildew, and leaf rust. The research process included data processing, increasing the data volume using augmentation techniques, model training, and evaluation of the results. Experimental results showed that the EfficientNet-B0 model produced the highest accuracy of 97.2%, significantly better than the CNN model built from scratch with an accuracy of 85.1%. This study demonstrates that transfer learning is highly effective in improving the accuracy of flower disease detection, making it a more reliable alternative to methods that do not utilize pre-trained models, especially for agricultural applications that require high levels of accuracy in disease detection.