Tomato crop yields face significant threats from plant diseases, with existing deep learning solutions often computationally prohibitive for resource-constrained agricultural settings; to address this gap, we propose Efficient Disease Attention Network (EDANet), a novel lightweight architecture combining depthwise separable convolutions with hybrid attention mechanisms for efficient Tomato disease recognition. Our approach integrates channel and spatial attention within hierarchical blocks to prioritize symptomatic regions while utilizing depthwise decomposition to reduce parameters to only 104,043 (multiple times smaller than MobileNet and EfficientNet). Evaluated on ten tomato disease classes from PlantVillage, EDANet achieves 97.32% accuracy and exceptional (~1.00) micro-AUC, with perfect recognition of Mosaic virus (100% F1-score) and robust performance on challenging cases like Early blight (93.2% F1) and Target Spot (93.6% F1). The architecture processes 128×128 RGB images in ~23ms on standard CPUs, enabling real-time field diagnostics without GPU dependencies. This work bridges laboratory AI and practical farm deployment by optimizing the accuracy-efficiency tradeoff, providing farmers with an accessible tool for early disease intervention in resource-limited environments.

Face recognition (FR) is a widely used biometric technology. Nevertheless, achieving efficient and robust FR is still challenging due to variations in illumination, pose, and facial expression. A vital step in any FR system is to select the most informative features and eliminate the redundant ones. In this study, a hybrid approach combining Principal Component Analysis (PCA) and the Dolphin Swarm Algorithm (DSA) with Euclidean Distance as a lightweight classifier is proposed. Experiments were made by using the ORL dataset, which consists of 400 grayscale images. With a 98% recognition rate for this hybrid approach against a recognition rate of 90–92% that can be achieved by PCA only, the proposed PCA+DSA outperformed standalone PCA while still being computationally economical. The metrics of Recognition Rate, Receiver Operating Characteristic (ROC), Cumulative Match Curve (CMC), and Expected Performance Curve (EPC) provided numerous confirmations for this Hybrid model. Additionally, the convergence analysis corroborated DSA’s efficacy in feature selection as the fitness was nearly 92% after nine iterations. Without requiring sophisticated classifiers or deep learning models, our findings show that the identification rate can be improved by combining a bio-inspired optimization technique and the classical PCA method. However, the current study is limited to the ORL dataset in a controlled environment. Future research will focus on implementing and evaluating the system in real-time scenarios on larger and more diverse datasets to enhance its scalability and robustness in practical applications.

Metabolic syndrome (MetS) is influenced by parental traits and diet. Drosophila melanogaster is a potential disease model organism, sharing physiology and genetic similarities with humans. Previous research had focused on Drosophila melanogaster as a model organism for obesity and diabetes, but not for MetS. This research aimed to determine the effect of both parental diet and high sugar fat (HSF) intake on glucose, cholesterol, and triglyceride hemolymph levels of Drosophila melanogaster. Wild flies were purified in either control (standard) or MetS media (extra 3% sucrose and palm oil). Seventy-five pairs were divided into 5 groups, according to parental origin and feeding media, and maintained in five generations (F1-F5). Glucose, triglyceride, and cholesterol levels were measured using a colorimetric assay in three replications of each generation per group. Glucose, triglyceride, and cholesterol levels were significantly different in all treatment groups, the control groups, and between generations in each group (p<0.05). Higher glucose and triglyceride levels appeared in the youngest generation (F5) of all groups, and in the flies reared on HSF diets. Maternal HSF-exposed groups demonstrated a more pronounced impact of parental metabolic state on the glucose and triglyceride levels of the earlier generation. These findings highlight that parental exposure to HSF and prolonged HSF intake independently and synergistically lead to persistent and amplified metabolic dysregulation across generations. Drosophila melanogaster, modeled in this study, represents a novel experimental organism that is suitable for studying the epigenetic inheritance of MetS, gaining more consideration for the metabolic health consequences of long-term dietary habits and parental metabolic state.

The presence of natural antioxidants in medicinal plants plays a crucial role in inhibiting the detrimental effects of oxidative stress. The aim of this research is to explore more deeply all parts of A. esculenta L from flowers, fruits, seeds, leaves, and stems for the levels of compounds and antioxidant and enzyme inhibitor activities. The flowers demonstrated the highest TPC with 173.15942 ± 6.5083 mg GAE/g. The stems exhibited the lowest TPC value at 69.1967 ± 2.8408 mg GAE/g. The flowers also showed TFC value of 83.157 ± 2.021 mg QE/g while the stems displayed the lowest with 36.7240 ± 1.337 mg QE/g. IC50 value that the flowers possessed the highest antioxidant activity with 22.6539 ± 1.6452 mg/mL, whereas the stems displayed a slightly lower. In terms of the inhibitor of a-amylase activity, the flowers had an IC50 value of 102.4885 ± 11.4370 mg/mL whereas the stems had a lower. The highest IC50 value of the a-glucosidase inhibitor was 76.95 ± 12.0888 mg/mL in the flowers, and the lowest was in the stems. The highest IC50 of pancreatic lipase inhibitor was 109.5943 ± 9.7391 mg/mL in the flowers, and the lowest was in the stems. This study show that there is a relationship between the high content of total phenolic and total flavonoids on antioxidants, antidiabetic and antilipase activities.

Recognizing handwritten digits and mathematical symbols remains a nontrivial challenge due to handwriting variability and visual similarity among classes. While deep learning, particularly Convolutional Neural Networks (CNNs), has significantly advanced handwriting recognition, many existing solutions rely on deep, resource-intensive architectures. This study aims to develop a lightweight and efficient CNN model for multi-class classification of handwritten digits and mathematical symbols, with an emphasis on deployability in resource-constrained environments such as educational platforms and embedded systems. The proposed model, implemented in Julia using the Flux.jl library, features a compact architecture with only two convolutional layers and approximately 55,000 trainable parameters significantly smaller than typical deep CNNs. Trained and evaluated on a publicly available dataset of over 10,000 grayscale 28×28-pixel images across 19 symbol classes, the model achieves a test accuracy of 91.8% while maintaining low computational demands. This work contributes to the development of practical handwritten mathematical expression recognition systems and demonstrates the feasibility of using Julia for developing lightweight deep learning applications.
 
Keywords - Digits, Mathematical Symbol, Classification, CNN

Pengenalan wajah (face recognition) merupakan suatu pengembangan dari teknologi deteksi wajah. Pengenalan wajah manusia merupakan salah satu turunan dari sistem biometrik yang menggunakan pola wajah manusia sebagai objek identifikasi. Sistem tersebut menggunakan pola wajah manusia yang terdapat dalam sistem basis data sebagai penyimpanan, kemudian akan melakukan perbandingan dengan gambar yang akan diuji. Sistem pengenalan wajah memiliki beberapa kendala, seperti sulit untuk mengenali objek dengan tingkat pencahayaan berbeda pada saat proses pengambilan gambar. Untuk mengatasi permasalahan yang terjadi akibat variasi tingkat cahaya, dikembangkan perangkat lunak dengan menerapkan metode Singular Value Decomposition (SVD). Pada projek ini metode eigenface cukup baik dalam melakukan pengenalan wajah. Bahkan dengan ukuran foto wajah yang cukup kecil (48 × 48), metode ini masih mampu untuk mengenali wajah dua orang yang sama. Proses pelatihan dan pengujiannya juga relatif singkat. Teknik ini dinilai efektif dalam mengenali foto wajah dengan ukuran yang kecil dan jumlah yang banyak.
 
Kata Kunci - Dekomposisi Nilai Singular, Eigenface, Pengenalan Wajah

Emosi merupakan aspek penting dalam komunikasi manusia yang sering muncul melalui unggahan di media sosial. Emosi tersebut diekspresikan dalam teks berbahasa Indonesia di platform media sosial X. Penelitian ini bertujuan untuk mendeteksi lima kategori emosi, yaitu marah, takut, senang, cinta, dan sedih. Model yang digunakan adalah algoritma Long Short-Term Memory (LSTM) dengan representasi kata dari FastText. Model dilatih menggunakan metode EarlyStopping dan dievaluasi dengan metrik akurasi, presisi, recall, dan F1-score. Hasil  menunjukkan bahwa model mencapai akurasi sebesar 79% pada data testing dengan performa yang relatif seimbang untuk setiap kategori emosi. Penelitian ini menunjukkan bahwa FastText dan LSTM efektif untuk mendeteksi emosi dalam teks media sosial berbahasa Indonesia. Penelitian ini diharapkan bermanfaat  dalam pengembangan penelitian berbasis emosi, seperti analisis sentimen, pemantauan opini publik, dan sistem pendukung kesehatan mental.
 
Kata Kunci – Deteksi Emosi, Sosial Media, Long Short-Term Memory, FastText

Accurate demand forecasting plays a vital role in optimizing inventory and distribution planning, especially for perishable goods such as bread. This study develops a time series forecasting model using a Recurrent Neural Network (RNN) with a Sequential architecture to predict daily bread demand. Unlike previous research, this model is trained on two years of real operational delivery data (2023–2024), enabling it to capture actual consumption patterns more effectively. The model leverages a 7-day sequence window to predict the next day’s demand, reflecting weekly seasonality. Data preprocessing includes normalization and cleaning, followed by training with the Stochastic Gradient Descent (SGD) optimizer. The model achieved a Mean Absolute Percentage Error (MAPE) of 4.88% and an accuracy of 86.90%, demonstrating high predictive performance and robustness in handling fluctuating, real-world data. The implementation of this model provides a practical solution for improving production planning, reducing waste, and enhancing supply chain responsiveness. The findings confirm that RNN-based models are effective tools for demand forecasting in dynamic business environments.
 
Keywords - Forecasting, Recurrent Neural Network (RNN), Demand Prediction, Operational Delivery Data, Bread Industry.

The Cargo oil pump is one of the auxiliary units on board a ship, used to transfer cargo from one location to another during loading and unloading operations. On the MT?Pangrango vessel, an issue occurred with the cargo oil pump: its performance declined, resulting in slow suction during the unloading process. This problem hindered efficient cargo operations, as the pump was not functioning properly.This study aims to identify the contributing factors, impacts, and countermeasures for the decreased performance of the cargo oil pump. The research methodology applied in this thesis involved a descriptive approach combined with Miles and Huberman’s analytical method, with data validity tested via triangulation. Data collection techniques used to identify the problem included observation, interviews, documentation, and literature review.The results revealed that the pump’s shaft was worn due to excessive friction, and the ball bearing s exhibited wear and disintegration caused by lack of lubrication (grease). The recommended corrective actions were to add material to the shaft , perform realignment to correct any shaft misalignment, and implement routine maintenance—specifically, monthly grease lubrication. These measures, when properly applied, can significantly reduce the risk of Cargo oil pump damage.

Penelitian ini mengkaji dinamika akses dan kontestasi kepentingan aktor dalam perebutan lokasi pertambangan rakyat di Provinsi Jambi dan Bangka Belitung. Fokus kajian meliputi kondisi lapangan, faktor ekonomi yang mendorong partisipasi masyarakat, serta relasi kekuasaan antar aktor. Pendekatan kualitatif diterapkan dengan pengumpulan data melalui observasi, wawancara purposif, dan studi dokumentasi. Hasil menunjukkan pertambangan rakyat berkembang pesat karena tingginya potensi sumber daya alam dan kebutuhan ekonomi masyarakat, meskipun sebagian besar kegiatan masih berstatus ilegal. Ketidaksesuaian regulasi dan praktik perizinan menciptakan ketidakpastian hukum yang mengancam keberlanjutan aktivitas tersebut. Pertambangan rakyat menjadi sumber utama penghidupan dengan dampak positif pada pendapatan dan kualitas hidup, namun juga menimbulkan tantangan keberlanjutan, ketimpangan sosial-ekonomi, serta risiko lingkungan. Kontestasi kepentingan memperlihatkan konflik struktural antara masyarakat, elite politik, dan pemilik modal yang didorong oleh lemahnya tata kelola dan regulasi. Dominasi perusahaan besar dan keterlibatan aktor lokal dalam praktik rente memperkuat perebutan sumber daya sebagai arena kekuasaan. Studi lanjutan disarankan untuk mengkaji narasi dan wacana antar pemangku kepentingan guna memperdalam pemahaman pembentukan kebijakan pertambangan rakyat di tingkat lokal.This study examines the dynamics of access and the contestation of interests among actors in the struggle over artisanal mining locations in the Jambi and Bangka Belitung Provinces. The focus of the study includes field conditions, economic factors driving community participation, and power relations among actors. A qualitative approach is applied, with data collected through observation, purposive interviews, and document studies. The results show that artisanal mining is rapidly developing due to the high potential of natural resources and the economic needs of the community, although most activities remain illegal. Regulatory mismatches and licensing practices create legal uncertainty that threatens the sustainability of these activities. Artisanal mining serves as a primary livelihood source with positive impacts on income and quality of life, but it also poses challenges related to sustainability, socio-economic inequality, and environmental risks. The contestation of interests reveals structural conflicts between communities, political elites, and capital owners, driven by weak governance and regulation. The dominance of large companies and the involvement of local actors in rent-seeking practices strengthen the competition over resources as a power arena. Further studies are recommended to examine the narratives and discourses among stakeholders to deepen the understanding of artisanal mining policy formation at the local level.