From prompts to packets: A view from the network on ChatGPT, Copilot, and Gemini

This study aims to analyse the distribution chain and consumption levels of wild-caught and farmed fish in the district of Yamoussoukro and the Guessabo sub-prefecture. To this end, the study will be based on documentary research, direct field observation and a questionnaire survey designed using KoboToolbox software. These questionnaires were administered to fishermen, fish farmers, restaurant owners and consumers in the aforementioned localities using Android mobile phones. The field survey was conducted over a period of five months, from September to January. The collected data was analysed using Excel and Word software to produce tables and graphs. The survey results reveal that, in Yamoussoukro and Guessabo, the most commonly produced and marketed fish species in fish farming is tilapia (Oreochromis niloticus). In the natural environment, however, the two main species of fish produced are tilapia and Mâchoiron. Fishermen and fish farmers use three types of distribution channel to market their products: direct channels, short-cycle intermediate channels and long-cycle intermediate channels. Specifically, fishermen focus on long-cycle intermediate channels and direct and short-cycle intermediate channels. The products are partly marketed in the production areas, but also in other cities to increase turnover. Most consumers prefer wild-caught fish based on taste and freshness criteria. It should also be noted that there has been a decline in fish stocks due to poor fishing practices and the impact of human activities in the surrounding area, such as agriculture and gold mining.

Early detection of wood-boring pest damage is crucial for preserving forest ecosystems and preventing economic losses in the timber industry. To address the challenges of extremely small early-stage damage symptoms, complex bark backgrounds, and deployment constraints on edge devices in forestry environments, this study proposes a light-weight non-intrusive detection framework, termed BH-YOLO, for early-stage damage symptom detection of wood-boring pests in trees. Here, "non-intrusive" refers to a fully contact-free and non-destructive detection paradigm that relies solely on RGB image acquisition and avoids invasive sensing methods such as drilling or sensor embedding, thereby balancing field diagnostic accuracy and edge-side deployment efficiency. Unlike conventional detectors that sacrifice accuracy for lightweight design, BH-YOLO achieves a superior accuracy-speed trade-off, making it particularly suitable for real-time deployment on resource-constrained edge devices. First, a Multi-Scale Ghost Convolution (MSGConv) module is integrated into the backbone to replace standard convolutions, leveraging parallel multi-scale kernel fusion to augment the model's perceptive capacity for irregular boreholes while effectively reducing topological complexity. Second, a Multi-Scale Attention Module with Channel Shuffle (MSAMCS) module is designed to supersede the conventional Bottleneck, significantly strengthening the extraction of core semantic features within complex textural environments. To further capture cryptic infestation symptoms, an Efficient Multi-scale Channel Attention (EMCA) module is embedded following the SPPF module, enhancing discrimination sensitivity toward minuscule targets through dynamic weight redistribution. Finally, Depthwise Separable Convolutions (DSConv) are employed to systematically replace the standard convolution layers and C2f modules in the neck network, achieving a drastic compression of parameter count and computational overhead. Experimental results demonstrate that BH-YOLO achieves an mAP@0.5 of 92.2% on the constructed real-world agricultural scene dataset, representing a 3.2% improvement over the baseline model, while the F1-score reaches 90.2%. In addition, the model's parameter count, floating-point operations, and model size are reduced by 66.7%, 64.6%, and 64.5%, respectively, while achieving an inference speed of 259.2 FPS for real-time mobile deployment-demonstrating its strong potential for edge-side deployment in real-world forestry monitoring scenarios. The proposed BH-YOLO model has been successfully deployed on Android mobile terminals, demonstrating robust performance across diverse illumination conditions and tree species, and specifically addresses the challenge of detecting extremely small, cryptic early-stage boreholes against complex bark textures, providing an efficient and automated solution for early monitoring and control of wood-boring pests while supporting forest ecosystem protection and reducing ecological and economic losses.

Despite the growing integration of technology in mathematics education, research on mobile learning platforms embedding local cultural contexts—particularly ethnomathematics—remains limited at the junior secondary level in Indonesia. Mathematical problem-solving ability remains a persistent challenge among students, necessitating Android-based learning media as a more engaging instructional solution. This investigation examined the extent to which an Android-based interactive mobile learning (m-learning) platform, integrated with ethnomathematics derived from Pontianak's local culture. A quasi-experimental posttest-only control-group design was adopted for this purpose. From nine classes at SMP Negeri 2 Pontianak, 60 participants were selected via cluster random sampling—30 in the experimental group receiving m-learning instruction and 30 in the comparison group receiving instruction without m-learning. Assessment of mathematical problem-solving capacity was conducted through a descriptive test comprising three items structured around Polya's four-stage indicators, with empirically established validity coefficients ranging from 0.94 to 0.96 and a reliability coefficient of 0.78. Statistical examination involved a one-tailed independent samples t-test for group comparison, supplemented by Cohen's d for effect size quantification. Findings revealed that learners exposed to m-learning instruction demonstrated substantially superior mathematical problem-solving performance compared to their counterparts in the conventional group (mean = 84.67 vs 72.63; p < 0.05), with a large effect magnitude (Cohen's d = 1.3). These outcomes affirm that the ethnomathematics-embedded Android-based m-learning platform constitutes an efficacious and educationally meaningful intervention for cultivating students' mathematical problem-solving proficiency. Nevertheless, further research on a broader scale is warranted to enhance the generalizability of these findings, along with the incorporation of a pretest to more rigorously assess gains and establish stronger causal evidence of the intervention's effectiveness.

The rapid expansion of mobile banking in emerging economies has increased exposure to client-side security risks, while MASVS-based security maturity benchmarking of conventional banking applications remains underrepresented in the literature. This study conducts a standard-based comparative security maturity assessment of two major Indonesian Android banking applications, BRImo and myBCA. APK files obtained from the Google Play Store were analysed using Static Application Security Testing with the Mobile Security Framework (MobSF) and evaluated against OWASP MASVS Level 2 and MASVS-R. MobSF scores were interpreted as relative indicators of security maturity based on severity-weighted findings across multiple domains. The results reveal a clear divergence in maturity levels. Although both applications demonstrate strong network-layer protection, BRImo exhibits structural weaknesses in storage, cryptography, platform interaction, and resilience domains, indicating fragmented defence-in-depth implementation. In contrast, myBCA shows more consistent cross-domain control integration. This study contributes an MASVS-based security maturity benchmarking approach and provides conceptual evidence that formal regulatory compliance may coexist with inconsistent client-side technical implementation. The findings offer analytically transferable insights for developers, security auditors, and regulators in rapidly digitalising financial ecosystems.

Resilience of forensic evidence acquisition under database schema drift

Recent advancements in Convolutional Neural Networks (CNNs), combined with the growing adoption of farm-applicable Internet of Things (IoT) devices, have expanded the application of precision agriculture in mango orchards. The Smart Mango Orchard can play a crucial role in ensuring mango trees thrive and produce high-quality fruit. However, state-of-the-art (SOTA) CNNs are built on numerous layers and many parameters; therefore, they are challenging to deploy in IoT devices. However, the lightweight CNN is a possible solution. This study developed a lightweight CNN, mangoNet, to deploy in an innovative mango orchard environment. The mangoNet is expected to monitor the mango leaf images and report them to farmers via a mobile app using the IoT system. The study was conducted using the primary dataset collected from the mango gardens in Rajshahi, Bangladesh. The mangoNet benchmark was evaluated using six SOTA CNNs. The mangoNet, with only 3,987,400 Parameters, outperforms SOTA CNNs’ accuracy (99.61%). In addition, this study employed SHAP, LIME, and Grad-CAM visualizations to identify and depict the image regions that contribute to mangoNet’s decision-making process. The mangoNet is integrated into a Streamlit web application and an Android mobile app, as researchers suggest for the practical use of CNNs. The novelty of mangoNet lies in its balanced sequential architecture, with a careful selection of kernels and progressive filter expansion, enabling early layers to capture low-level features and deeper layers to extract high-level features. As a result, the proposed mangoNet achieved high accuracy while requiring fewer computational resources and reduced training time. In addition, the mangoNet-powered website and mobile application empower both farmers and farming stakeholders by making real-time disease detection. In the future, the prototype is expected to be commercialized as Bangladesh is the 8Th mango-producing country in the world.

The rise of cyber-attacks targeting mobile devices, particularly Android malware, has continued to grow, making the development of automated detection and classification methods increasingly important. This study aims to compare the performance of five supervised learning algorithms, namely Decision Tree (DT), Random Forest (RF), Support Vector Machine (SVM), K-Nearest Neighbors (KNN), and Naïve Bayes (NB), in detecting and classifying Android malware applications. The experiments were conducted using the CICMalDroid2020 dataset, which consists of multiple malware categories as well as benign applications. Feature selection procedure was implemented, approach was designed to accelerate training and improve predictive accuracy by retaining only most relevant predictors. All algorithms were trained with commonly employed default hyperparameters to establish a fair baseline rather than conducting extensive parameter optimisation. Two evaluation strategies were employed: 10-fold cross-validation, and holdout method with 70/30 training split. The performance was assessed using standard metrics including Accuracy, Precision, Recall, and F1-score. The experimental results indicate that RF consistently achieved the highest performance across both evaluation methods 94.17% Accuracy with 10-fold cross-validation, 93.65% Accuracy with the holdout split. In contrast, NB showed the lowest performance in all metrics, while DT and KNN delivered relatively competitive results with acceptable accuracy. SVM, however, produced lower accuracy compared to RF and DT. These findings highlight importance of feature selection and significance of selecting an appropriate algorithm in Android malware detection. Although RF demonstrated robustness on this large, complex dataset, further research is required to assess its computational cost and scalability for deployment on resource-constrained mobile devices.

The rate of increasing cleanliness issues has been considered as a biggest problem in today’s modern scenario. The issue has created a serious impact on the overall well-being of many citizens in both rural and urban areas. The traditional checking of cleanliness methods needs manual checking, which results in poor progress and no effective methodology. Since India is a vast country, checking each place manually is an impossible act. To overcome these drawbacks, we have introduced a software solution that uses CNN to automatically detect and report unclean public places. In regard to this, the system uses an Android mobile to take pictures of public areas. So, this would be simpler, and there is no special equipment needed. Here, these photos are fed into a CNN model that uses the patterns to identify trash. The system generates a report with the photo, GPS location, and time details when it detects a dirty area. The municipal authorities can receive the report directly. The system keeps these records for a while, so the authorities can track the patterns and identify problems in cleanliness maintenance. As per this project, mobile data collection and CNN analysis are combined regarding urban cleanliness management. This can be a very efficient solution for cities. It reduces human work and improves detection. Further, this system helps in quick action for the maintenance of clean and healthier cities. When a person uploads a photo, they will get some rewards. This will make the work more efficient. By this approach, we can conclude that if this is made as a daily technology, then it becomes easier to solve practical problems in city management.

The agriculture and aquaculture sector has adopted the concept of smart farming based on the Internet of Things (IoT) to monitor and control the cultivation environment with precision. However, the efficiency achieved in the pre-harvest phase is often stopped in the post-harvest phase, especially in yield counting activities. The manual calculation process is highly inefficient and prone to human error, which has a direct impact on logistics management and agribusiness profitability. To overcome these inefficiencies, this study proposes the design and implementation of an automated crop yield counting system that integrates Image Processing and IoT technologies. The system is designed to use cameras and image processing algorithms to accurately detect and calculate the quantity of crops. The calculated data is sent to the IoT cloud platform for real-time storage and access through the Android mobile app. The hardware framework includes microcontrollers such as the ESP32/ESP8266, which have proven reliable in other IoT monitoring systems. The expected result of this study is a prototype system that is able to calculate crop yields with a high level of accuracy and speed that far exceeds manual methods. The Android app implementation provides an intuitive monitoring and control interface, allowing users to access historical and real-time data at any time. The main contribution of this research is to bridge the digitalization gap in the post-harvest phase, as well as provide fast, accurate, and digitized solutions to support comprehensive precision agriculture management.

Abstract. Technological advances in virtual reality (VR) in recent years have the potential to impact our everyday lives increasingly. VR enables us to explore a digital world through an immersive experience with a head-mounted display (HMD). When combined with tools for 3D documentation and modelling, as well as software for creating interactive virtual worlds, VR can play a significant role in preserving and visualising cultural heritage for museums, educational institutions, and other cultural sectors. VR opens up a new form of scientific communication that can benefit damaged, destroyed or distant historical and cultural heritage objects. On the other hand, augmented reality (AR) can resurrect and display forgotten and lost cultural heritage objects and monuments in their original locations via mobile devices. This paper presents the development of a VR and AR application of the disappeared Horneburg castle, which does not exist anymore since almost 400 years. These applications will serve as tourist attractions for visitors to the municipality of Horneburg, enabling them to experience the castle's historic life in an immersive way at the museum and interactively on site. Based on the uncovering of the underground remains of the castle by geomagnetic prospecting the Horneburg Castle was reconstructed using a derived 2D site plan and architectural drawings of the castle buildings. The 3D models of the buildings were constructed using the 2D drawings, which were then transferred to the game engine Unreal 5 for texture mapping and immersive visualisation. The AR app was developed to visualise the whole building ensemble with the wooden palisade around the castle or each individual building on mobile android and iOS devices using WebXR and Three.js on site in nature (option 1) and location-independent (option 2). The technical implementation of the VR and AR applications is presented in this article.

The manual school payment system creates quite long queues and does not create time efficiency. Parents of students are required to come to school, especially to the Administration section because the system is still manual. This research was conducted at a Private Elementary School in Bekasi City. The split payment system makes many parents confused and often scattered. Every month there are several fees that parents must pay at this school. Tuition fees, Parent Teacher Association (POMG), Teaching and Learning Activities (KBM), and Books. It is hoped that this research will make it easier for parents so that the financial system in the school can run well and according to expectations. Then there is no longer a long queue system because the administration counter is only filled by 1 school employee. So that every parent can make school payments in real time, 24 hours whenever and wherever just by accessing this application. Of course, this payment application works with a bank in collaboration with the school foundation.

User interactions with mobile applications (apps) are accompanied by continuous visual changes in the Graphical User Interface (GUI), guiding task completion and feedback. These changes help users complete intended tasks or assess the appropriateness of their actions, typically conveyed through visual cues such as appearance and color. While such visual changes are effective for sighted users, they are inaccessible to blind users, creating substantial barriers to GUI interaction. To address these challenges, we propose VisualDroid , a method based on a multi-modal large language model (LLM) for testing and classifying GUI visual changes using a tailored three-hop reasoning prompting framework. VisualDroid achieved an F1-score of 94.7% in 34 apps from 17 domains, surpassing all baseline methods. When evaluated on five open-source apps from F-Droid, our method enabled developers to resolve three identified issues, with two still under review. In terms of efficiency and cost, our method indicates minimal resource consumption.

In this paper, we have experimenting to identifying the Kannada emotional speech signals recognition using deep learning technique. Speech Emotion Recognition (SER) is a hot research topic in the field of HCI (Human Computer Interaction). To experiment we have created our own Kannada Emotional Speech corpus recording in different android mobile phone, with sampling rate 48 kbps and mono channel and speakers are from the Karnataka and Tamil Nadu boarder region, whose accents is a mixer of Kannada, Tamil and Telugu languages. The emotions are anger, fear, sad and neutral is considered. Totally we have 440 Kannada emotional speech signals. To use this speech signal data in our experiment, we have first down sampled all speech signals into 16 kbps, mono channel and stored as wav files. Then we pass the each sample signals into the pre-emphasize phase, then framing and windowing is done, then we extracted features from each voice frame signal like MFCC, Chroma and ‘Mel’ and then passed to CNN (Convolution Neural Network) to train and testing the voice signals. The CNN is capable to handle small amount of data for training, and with smaller number of parameters. The average Kannada emotion recognition accuracy rate of 93.77% is achieved. All computations are done using Python programming language and used Python libraries like Librosa, Keras, pyaudio, soundfile, sklearn.

Pre-exposure prophylaxis (PrEP) is a highly effective biomedical prevention tool for HIV yet remains underutilized among key populations, particularly among young sexual and gender minorities (SGM). Recognizing the popularity of specific dating and social media apps among SGM young adults, we leveraged user data from these platforms to build a machine learning (ML) model that could inform targeted, data-driven interventions aimed at improving PrEP uptake and adherence. We adapted eWellness, an Android mobile app, to passively collect data from research participants capturing mobile app usage, keystroke patterns and logs, and GPS location data between 2021 and 2024. These data were used to train a ML model to predict self-reported PrEP use. Model accuracy was evaluated through F1 scores across different data types and feature combinations. Study protocols were developed in collaboration with community partners and adhered to strict ethical and privacy standards. A total of 82 SGM young adults participated, with 46 (56%) reporting PrEP use at baseline. Our machine learning model demonstrated good predictive accuracy for predicting PrEP use and non-use, achieving an F1 score of 0.84 (PrEP use) and 0.82 (non-use) outcomes when incorporating data from all mobile apps, including messaging, dating, and social media mobile apps. By contrast, predictions based solely on social media mobile app usage, language associated with sexual behavior and substance use risk, or location monitoring demonstrated worse accuracy (F1 scores of 0.79/0.75, 0.70/0.57, and 0.70/0.52, respectively). Additional feature extraction methods, as well as various combinations of these features, were also tested. However, none achieved predictive accuracy as well as the model incorporating all mobile app usage data combined. This study demonstrates the potential of machine learning to accurately predict PrEP use status among SGM young adults. The findings offer a foundation for developing more personalized PrEP promotion strategies, particularly among SGM young adults who use social media and dating apps. Future research should assess the model's adaptability across diverse SGM subgroups to further inform intervention development. Registry: ClinicalTrials.gov, ID: NCT04710901, November 9, 2020.

The purpose of this study is to determine the influence of cultural factors and personal factors on consumer satisfaction in purchasing Android mobile phones in the Sliwung Stubondo Village Community. The data results in this study were obtained by distributing questionnaires to respondents who use Android mobile phone products. The population in this study is the Sliwung Village community. Sampling in this study used purposive sampling with proportional random sampling techniques and data analysis used was multiple linear regression. The calculated F value is 27.180 where this value is a test statistic that compares with the F table for α = 0.05 with a degree of 2 and a degree of mention of 58, then the F table value is 3.15. The calculated F ≥ F table (27.180 ≥ 3.15), in addition there is also a significant value of 0.01 so it is smaller than α = 0.05. Based on a simultaneous F-test, cultural and personal factors were significant (27.180 ≥ 3.15), indicating that both cultural and personal factors influence consumer satisfaction when purchasing an Android phone.

Purpose. To conduct a systematic analysis of available mobile applications for Android and iOS designed for performing practical tasks during introductory, geological mapping, and specialized field practices as well as detailed geological investigations; to assess their capabilities as portable digital tools (advantages and limitations) for increasing the efficiency of fieldwork of various types and levels of complexity; and to identify the most functional solutions for collecting spatially referenced data, recording specialized geological information, digital geological mapping, and structural measurements. Methodology. Digital geological field data were collected, systematically recorded, and visualized (including data for geological mapping and specialized measurements) using mobile software tools for Android and iOS in combination with built-in smartphone sensors (GPS / GLONASS / GNSS, accelerometer, gyroscope, magnetometer, barometer, etc.). The data sources included direct field observations, results of structural measurements, georeferenced photographic documentation, and other geodata acquired in accordance with the research objectives. Results. The functional capabilities, tools and features of the use of mobile geological applications for smartphones on the Android (Google Play) and iOS (Apple App Store) platforms were analyzed. Approaches to using these applications as means of input, storage and primary analysis of structural and related geospatial data in field conditions were shown. Using the example of several representative applications — Geology Clinometer: GeoCompass, SW Maps, FieldMove Clino, QField for QGIS — the possibility of performing a full cycle of structural digital data collection using smartphones, as well as their subsequent export to desktop GIS platforms on personal computers and laptops for detailed processing and interpretation, was demonstrated. The results obtained confirm the suitability of mobile solutions as effective portable tools for field geological research of various levels of complexity. Scientific novelty. The rapid advancement of smartphone hardware, along with the emergence and evolution of specialized mobile applications for geological tasks, necessitates a revision of traditional approaches to organizing field investigations. It is shown that mobile applications, when combined with built-in sensors of modern smartphones, can provide not only rapid acquisition of geodata, but also their immediate visualization, preliminary analysis, and structured digital storage without the use of bulky field instruments. This demonstrates a transition from analog and paper-based methods to portable digital working environments, which radically increases the speed and reproducibility of field operations and opens new possibilities for the standardization and integration of field data into subsequent GIS-based analysis. Thus, the study confirms the methodological transformation of field geology under the influence of mobile technologies as a qualitatively new tool in the geosciences. Practical significance. The introduction of mobile applications into field geological practice directly increases the quality and reproducibility of field investigations. By enabling rapid and reliable data entry, structured storage, and immediate visualization of geospatial information, these tools provide a robust foundation for digital geological mapping and subsequent analytical processing. The integration of Android and iOS mobile applications with GIS platforms (QGIS, ArcGIS Pro) allows data from different field stages to be combined, compared, updated, and incorporated into high-accuracy digital maps. Thus, the use of mobile applications substantially enhances the productivity, reliability, and standardization of modern field-based geological research.

As a compulsory course for software engineering and IoT engineering majors, "Android Mobile Application Development" focuses on Android core components, data storage, and map applications, aiming to cultivate students' ability to develop Android applications for IoT terminals. In response to the common problems in current computer education, such as excessive emphasis on theoretical courses, insufficient cultivation of practical abilities, and a disconnect between graduates' skills and the needs of enterprises. This article proposes a project-based learning model that can effectively stimulate students' autonomy, innovative thinking, teamwork, and engineering practice abilities by guiding them to solve real and complex engineering projects, which is in line with the goal of cultivating applied talents.

The increasing use of Android mobile devices and applications leads to an increase in malware threats. There is a requirement to investigate if a more detailed feature extraction from APK files with deep learning can produce more accurate results. We investigate using deep learning techniques to detect Android Malware considering the latest datasets. We aim to improve the system’s ability to accurately classify and detect a wider range of Android malware variants. We propose a mechanism to carry out APK analysis for feature extraction capable of extracting 46,648 features. We retain 10,523 features after applying feature selection and subsequently use these selected features to train the neural networks. We make use of APK retrieved from Androzoo for dataset generation. We contribute a dataset with code and scripts to arrive at our proposed dataset using a public repository. We compare deep learning models based on deep neural networks (DNN), convolutional neural networks (CNN), and transfer learning-based models using static features. We consider our contributed datasets and conclude that the DNN-based models outperform the CNN models with a wider range and number of features.

Android employs Ahead-of-Time (AOT) pre-compilation to enhance application launch speed and runtime performance. However, these generated OAT files over-consume the scarce memory and storage resources on mobile devices, leading to degraded user experience. Our analysis of several Android applications reveals an average code redundancy of 25.4%. To reduce the code size via redundancy elimination, we propose Calibro , a C ompilation- a ssisted li nk-time b ina r y code o utlining method. However, it will incur high build overhead, so we introduce several optimizations to better suit resource-limited mobile devices. Besides, we also propose optional filtering strategies to further meet performance requirements. Experimental results show that, under common scenarios, our method reduces the OAT file code size by 19.6% and the runtime memory usage by 15.4% on average, with negligible runtime performance degradation in terms of user experience and tolerable build overhead. Therefore, the proposed method shows promise for industrial deployment on real-world Android mobile devices.