Classical machine learning (ML) models, including Logistic Regression (LR), Support Vector Machines (SVM), Random Forests (RF), and XGBoost, remain widely used in practical applications because of their efficiency, interpretability, and relatively low computational costs. However, their security properties against different adversarial threats are often evaluated independently rather than within a unified comparative framework. This paper presents a unified empirical evaluation of Logistic Regression, Linear SVM, Random Forest, and XGBoost models across image (MNIST and CIFAR-10), text (AG News), and tabular (Adult, Spambase) classification domains. The models are evaluated under three adversarial scenarios: training-time poisoning, inference-time evasion, and black-box model extraction attacks. The empirical results reveal architecture- and attack-dependent security trade-offs. Random Forest demonstrated relatively greater resilience to random label noise in several configurations but remained vulnerable to targeted poisoning, particularly on AG News. In contrast, Logistic Regression and Linear SVM achieved high black-box extraction fidelity under some active-query configurations, reaching 96.08% and 94.02%, respectively, on MNIST with Q = 10,000. Their evasion results were comparatively interpretable under the evaluated attack procedures, although the observed performance depended on the perturbation budget and optimization configuration. In particular, the higher apparent accuracy of Linear SVM under PGD than under FGSM should be interpreted as a configuration-dependent observation rather than evidence of intrinsic robustness or confirmed gradient masking. Overall, the findings indicate that model robustness depends on the interaction between the attack strategy, data representation, and model architecture.

Illegal waste dumping detection in surveillance video is fundamentally a temporal novelty problem: the goal is to detect the moment when new waste appears, not merely to identify all waste visible in a scene. We propose a hybrid real-time pipeline that combines (i) a YOLOv8m detector operated at low confidence to maximize recall, (ii) persistent multi-object tracking with BoT-SORT to build temporal object histories, and (iii) a multi-gate novelty validator that fuses complementary cues-area change, motion dynamics, dual-timescale background models, and optical flow-to distinguish newly deposited objects from pre-existing clutter. A temporal confirmation stage enforces stationarity and persistence criteria before reporting the precise dumping timestamp. Evaluated on the private test set of the Illegal Waste Dumping Detection (IWDD) 2026 contest, our system achieves an F1-score of 0.40 (Precision: 0.43, Recall: 0.38) and a Detection-Norm (D-Norm) score of 0.78, indicating strong temporal alignment with ground truth events. The high D-Norm relative to F1 demonstrates that the hybrid architecture successfully identifies when dumping occurs with precision, while false negatives stem primarily from challenging visual conditions rather than temporal misalignment. These results confirm the system's capability for robust event detection in diverse surveillance scenarios while maintaining computational efficiency suitable for edge deployment.

The rapid advancement of quantum computing poses a fundamental threat to classical public-key cryptographic systems, necessitating the transition to post-quantum cryptography (PQC). While significant progress has been made in the standardization of quantum-resistant algorithms, their practical deployment in heterogeneous environments—particularly resource-constrained Internet of Things (IoT) devices—remains a critical challenge. This study presents a comprehensive experimental evaluation of four NIST-standardized PQC algorithms: CRYSTALS-Kyber (ML-KEM), CRYSTALS-Dilithium (ML-DSA), FALCON, and SPHINCS+. The scope of these findings is bounded by an empirical analysis conducted across two specific testing platforms, a high-performance x86-64 workstation (AMD Ryzen 7 5700U) and a resource-constrained embedded microcontroller (ESP32-WROOM), utilizing dedicated software environments implemented in Native C, Go, and Python. The evaluation isolates key performance indicators, including computational latency, memory consumption, communication overhead, and temporal determinism, based on benchmarking over 1000 iterations. Within this experimental setup, results demonstrate clear trade-offs between target security categories, execution performance, and structural memory limits. Lattice-based schemes such as Kyber and Falcon exhibit optimal efficiency and scalability on the tested embedded platform, while the specific memory limits of the ESP32 platform introduce architectural stability constraints for higher-tier Dilithium variants. In contrast, SPHINCS+ provides structural robustness at the cost of higher computational hashing latency within these evaluation environments. The findings highlight the critical role of hardware-specific constraints and language runtime design choices in enabling practical PQC deployment, providing context-specific insights supporting the secure migration of IoT infrastructures toward quantum-resilient systems.

With the proliferation of IoT-based applications, security requirements are becoming increasingly stringent. Given the diversity of such systems, selecting the most appropriate solutions and technologies to address the challenges is a complex activity. This paper provides an exhaustive evaluation of existing security challenges related to the IoT domain, analysing studies published between 2021 and 2025. This review explores the evolving landscape of IoT security, identifying key focus areas, challenges, and proposed solutions as presented in recent research. Through this analysis, the review categorizes IoT security efforts into six main areas: emerging technologies (35.2% of studies), securing identity management (19.3%), attack detection (17.9%), data management and protection (8.3%), communication and networking (13.8%), and risk management (5.5%). These percentages highlight the research community’s focus and indicate areas requiring further investigation. From leveraging machine learning and blockchain for anomaly detection and real-time threat response to optimising lightweight algorithms for resource-limited devices, researchers propose innovative and adaptive solutions to address emerging threats. The review underscores the integration of advanced technologies to enhance IoT system security, while also highlighting ongoing challenges. The paper concludes with a synthesis of security challenges and threats of each identified category, along with their solutions, aiming to support decision-making during the design approach of IoT-based applications and to guide future research toward comprehensive and efficient IoT frameworks.

Blockchain technology has rapidly evolved as a cornerstone of decentralized computing, transforming how trust, data integrity, and transparency are achieved in digital ecosystems. However, despite extensive adoption, significant gaps remain in understanding how key blockchain variables, such as block size, consensus mechanisms, and network latency, affect system vulnerabilities and susceptibility to cyberattacks. This survey addresses this gap by combining qualitative and quantitative analyses across multiple blockchain environments. Using simulation tools such as Ganache and Bitcoin Core, and reviewing peer-reviewed studies from 2016 to 2024, the research systematically maps blockchain parameters to cyberattack vectors including 51% attacks, Sybil attacks, and double-spending. Findings indicate that design choices like block size, block interval, and consensus type substantially influence resilience against attacks. The Blockchain Variable Quantitative Risk Framework (BVQRF) introduced here integrates NIST’s cybersecurity principles with quantitative scoring to assess risks. This framework represents a novel contribution by operationalizing theoretical security constructs into actionable evaluation metrics, enabling predictive modeling and adaptive risk mitigation strategies for blockchain systems.

Diplomas and academic transcripts issued at the conclusion of a university cycle have been the subject of numerous studies focused on developing secure methods for their registration and access. However, in the context of high school teachers, these initial credentials mark only the starting point of a much more complex professional journey. Throughout their careers, teachers receive a wide array of certificates and attestations related to professional development, participation in educational projects, volunteering, and institutional contributions. Many of these documents are issued directly by the school administration and are often vulnerable to misplacement, unauthorized alterations, or limited portability. These challenges are amplified when teachers move between schools or are involved in teaching across multiple institutions. In response to this need, this paper proposes a blockchain-based solution built on the Ethereum platform, which ensures the integrity, traceability, and long-term accessibility of such records, preserving the professional achievements of teachers across their careers. Although most research has focused on securing highly valuable documents on blockchain, such as diplomas, certificates, and micro-credentials, this study highlights the importance of extending blockchain solutions to school-issued attestations, as they carry significant weight in teacher evaluation and the development of professional portfolios.

Blockchain governance has become central to the performance and resilience of decentralized systems, yet current models face recurring issues of participation, coordination, and adaptability. This article offers a structured analysis of governance frameworks and highlights their limitations through recent high-impact case studies. It then examines how artificial intelligence (AI) is being integrated into governance processes, ranging from proposal summarization and anomaly detection to autonomous agent-based voting. In response to existing gaps, this paper proposes the Voting Via Parallel Predictive Agents (VOPPA) framework, a multi-agent architecture aimed at enabling predictive, diverse, and decentralized decision-making. Strengthening blockchain governance will require not just decentralization but also intelligent, adaptable, and accountable decision-making systems.

Nowadays, the consequence of quick access to information has lead to the spread of fake news, which has a strong damaging impact on democracy, justice, and public trust. Thus, it is crucial to analyze and evaluate detection methods for fake news. This paper focuses on the detection of Romanian fake news. In this study, we made a comparative analysis of machine learning algorithms and Transformer-based models on Romanian fake news detection using three datasets—FakeRom, NEW, and both FakeRom + NEW. The NEW dataset was build using a scrapping algorithm applied on the Veridica platform. Our approach uses the following machine learning models for detection: Naive Bayes (NB), Logistic Regression (LR), and Support Vector Machine (SVM). We also used two Transformer-based models—BERT-based-multilingual-cased and RoBERTa-large. The performance of the models was evaluated using various metrics: accuracy, precision, recall, and F1 score. The results revealed that the BERT model trained on the NEW dataset consistently achieved the highest performance metrics across all test sets, with 96.5%. Also, Support Vector Machine trained on NEW was another top performer, reaching a very good accuracy of 94.6% on the combined test set.

The advancement of artificial intelligence (AI) technologies, including generative pre-trained transformers (GPTs) and generative models for text, image, audio, and video creation, has revolutionized content generation, creating unprecedented opportunities and critical challenges. This paper systematically examines the characteristics, methodologies, and challenges associated with detecting the synthetic content across multiple modalities, to safeguard digital authenticity and integrity. Key detection approaches reviewed include stylometric analysis, watermarking, pixel prediction techniques, dual-stream networks, machine learning models, blockchain, and hybrid approaches, highlighting their strengths and limitations, as well as their detection accuracy, independent accuracy of 80% for stylometric analysis and up to 92% using multiple modalities in hybrid approaches. The effectiveness of these techniques is explored in diverse contexts, from identifying deepfakes and synthetic media to detecting AI-generated scientific texts. Ethical concerns, such as privacy violations, algorithmic bias, false positives, and overreliance on automated systems, are also critically discussed. Furthermore, the paper addresses legal and regulatory frameworks, including intellectual property challenges and emerging legislation, emphasizing the need for robust governance to mitigate misuse. Real-world examples of detection systems are analyzed to provide practical insights into implementation challenges. Future directions include developing generalizable and adaptive detection models, hybrid approaches, fostering collaboration between stakeholders, and integrating ethical safeguards. By presenting a comprehensive overview of AIGC detection, this paper aims to inform stakeholders, researchers, policymakers, and practitioners on addressing the dual-edged implications of AI-driven content creation.

The quantity of textual data on the internet is growing exponentially, and it is very tough task to obtain important and relevant information from it. An efficient and effective method is required that provides a concise summary of an article. This can be achieved by the usage of automatic text summarization. In this research, the authors suggested an efficient approach for text summarization where an extractive summary is generated from an article. The methodology was modified by integrating a normalized similarity matrix of both BM25+ and conventional TextRank algorithm, which resulted in the improvised results. A graph is generated by taking the sentences in the article as nodes and edge weights as the similarity score between two sentences. The maximum rank nodes are selected, and the summary is extracted. Empirical evaluation of the proposed methodology was analyzed and compared with baseline methods viz. the conventional TextRank algorithm, term frequency–inverse document frequency (TF–IDF) cosine, longest common consequence (LCS), and BM25+ by taking precision, recall, and F1 score as evaluation criteria. ROUGE-1, ROUGE-2, and ROUGE-L scores were calculated for all the methods. The outcomes demonstrate that the proposed method can efficiently summarize any article irrespective of the category it belongs to.

Considering the problem of diabetes prediction for the use case of PIMA Indian Data Set and Diabetes Dataset 2019, the present paper compares the efficiency of two automatic machine learning tools without code offered by Microsoft and Matlab: Automated ML from Azure Machine Learning Studio, and the Classification Learner Application from Matlab. Also, the results obtained for the best model obtained using the automatic machine tools without code, are compared with those obtained for an implementation with code, made on-premise.

On university campuses, retrofitting studies have historically concentrated on individual buildings (or building components) instead of the entire campus. In the present paper, we examine how an incorporated strategic planning strategy might be used to investigate the socio-technical construction of a campus retrofit operation throughout multiple scale/sectors. The campus of the University of Oradea (CUO), Romania, with its beginnings in the 1910s was investigated using its new master plan. The developed strategies for a “green” and “healthy” campus depict a CUO redesign involving complex solutions for the green renovation of old buildings. In addition, the improvement effects of the modernization interventions already carried out were analyzed and quantified. Sixteen buildings (30% of the built area) were consolidated/rehabilitated/modernized, and/or equipped in the last decade, seven educational spaces being included in the circuit of the buildings fund (totaling 5491.59 sq m). For the renovated spaces, energy consumption was reduced by 20–88% and CO2 emissions by 41.82–86%, depending on the specifics of each space. The reconfiguration, rehabilitation, and energy efficiency of the entire heating system of CUO (which uses geothermal water as a specific characteristic) significantly improved (20% decrease in energy and 21% decrease in geothermal water consumption). Our findings offer new directions and design solutions for the ecological modernization of other outdated university campuses, highlighting new perspectives in the green university campuses’ management, as a way to implement sustainability in the higher education environment. Data presented give professionals in the field (architects, designers, engineers, planners, and decision makers) a clear picture of the benefits due to ecological renovation, also offering the necessary tools to implement new solutions for reducing the impact of urban areas on the environment.

Medical image analysis and classification is an important application of computer vision wherein disease prediction based on an input image is provided to assist healthcare professionals. There are many deep learning architectures that accept the different medical image modalities and provide the decisions about the diagnosis of various cancers, including breast cancer, cervical cancer, etc. The Pap-smear test is the commonly used diagnostic procedure for early identification of cervical cancer, but it has a high rate of false-positive results due to human error. Therefore, computer-aided diagnostic systems based on deep learning need to be further researched to classify the pap-smear images accurately. A fuzzy min–max neural network is a neuro fuzzy architecture that has many advantages, such as training with a minimum number of passes, handling overlapping class classification, supporting online training and adaptation, etc. This paper has proposed a novel hybrid technique that combines the deep learning architectures with machine learning classifiers and fuzzy min–max neural network for feature extraction and Pap-smear image classification, respectively. The deep learning pretrained models used are Alexnet, ResNet-18, ResNet-50, and GoogleNet. Benchmark datasets used for the experimentation are Herlev and Sipakmed. The highest classification accuracy of 95.33% is obtained using Resnet-50 fine-tuned architecture followed by Alexnet on Sipakmed dataset. In addition to the improved accuracies, the proposed model has utilized the advantages of fuzzy min–max neural network classifiers mentioned in the literature.

The presence of spam content in social media is tremendously increasing, and therefore the detection of spam has become vital. The spam contents increase as people extensively use social media, i.e., Facebook, Twitter, YouTube, and E-mail. The time spent by people using social media is overgrowing, especially in the time of the pandemic. Users get a lot of text messages through social media, and they cannot recognize the spam content in these messages. Spam messages contain malicious links, apps, fake accounts, fake news, reviews, rumors, etc. To improve social media security, the detection and control of spam text are essential. This paper presents a detailed survey on the latest developments in spam text detection and classification in social media. The various techniques involved in spam detection and classification involving Machine Learning, Deep Learning, and text-based approaches are discussed in this paper. We also present the challenges encountered in the identification of spam with its control mechanisms and datasets used in existing works involving spam detection.

The agricultural sector plays a key role in supplying quality food and makes the greatest contribution to growing economies and populations. Plant disease may cause significant losses in food production and eradicate diversity in species. Early diagnosis of plant diseases using accurate or automatic detection techniques can enhance the quality of food production and minimize economic losses. In recent years, deep learning has brought tremendous improvements in the recognition accuracy of image classification and object detection systems. Hence, in this paper, we utilized convolutional neural network (CNN)-based pre-trained models for efficient plant disease identification. We focused on fine tuning the hyperparameters of popular pre-trained models, such as DenseNet-121, ResNet-50, VGG-16, and Inception V4. The experiments were carried out using the popular PlantVillage dataset, which has 54,305 image samples of different plant disease species in 38 classes. The performance of the model was evaluated through classification accuracy, sensitivity, specificity, and F1 score. A comparative analysis was also performed with similar state-of-the-art studies. The experiments proved that DenseNet-121 achieved 99.81% higher classification accuracy, which was superior to state-of-the-art models.

Image captioning is oriented towards describing an image with the best possible use of words that can provide a semantic, relatable meaning of the scenario inscribed. Different models can be used to accomplish this arduous task depending on the context and requirement of what needs to be achieved. An encoder–decoder model which uses the image feature vectors as an input to the encoder is often marked as one of the appropriate models to accomplish the captioning process. In the proposed work, a dual-modal transformer has been used which captures the intra- and inter-model interactions in a simultaneous manner within an attention block. The transformer architecture is quantitatively evaluated on a publicly available Microsoft Common Objects in Context (MS COCO) dataset yielding a Bilingual Evaluation Understudy (BLEU)-4 Score of 85.01. The efficacy of the model is evaluated on Flickr 8k, Flickr 30k datasets and MS COCO datasets and results for the same is compared and analysed with the state-of-the-art methods. The results shows that the proposed model outperformed when compared with conventional models, such as the encoder–decoder model and attention model.

Examinations or assessments play a vital role in every student's life; they determine their future and career paths. The COVID pandemic has left adverse impacts in all areas, including the academic field. The regularized classroom learning and face-to-face real-time examinations were not feasible to avoid widespread infection and ensure safety. During these desperate times, technological advancements stepped in to aid students in continuing their education without any academic breaks. Machine learning is a key to this digital transformation of schools or colleges from real-time to online mode. Online learning and examination during lockdown were made possible by Machine learning methods. In this article, a systematic review of the role of Machine learning in Lockdown Exam Management Systems was conducted by evaluating 135 studies over the last five years. The significance of Machine learning in the entire exam cycle from pre-exam preparation, conduction of examination, and evaluation were studied and discussed. The unsupervised or supervised Machine learning algorithms were identified and categorized in each process. The primary aspects of examinations, such as authentication, scheduling, proctoring, and cheat or fraud detection, are investigated in detail with Machine learning perspectives. The main attributes, such as prediction of at-risk students, adaptive learning, and monitoring of students, are integrated for more understanding of the role of machine learning in exam preparation, followed by its management of the post-examination process. Finally, this review concludes with issues and challenges that machine learning imposes on the examination system, and these issues are discussed with solutions.

A novel image reversible data-hiding scheme based on primitive and varying radix numerical model is presented in this article. Using varying radix, variable sum of data may be embedded in various pixels of images. This scheme is made adaptive using the correlation of the neighboring pixels. Messages are embedded as blocks of non-uniform length in the high-frequency regions of the rhombus mean interpolated image. A higher amount of data is embedded in the high-frequency regions and lesser data in the low-frequency regions of the image. The size of the embedded data depends on the statistics of the pixel distribution in the cover image. One of the major issues in reversible data embedding, the location map, is minimized because of the interpolation process. This technique, which is actually LSB matching, embeds only the residuals of modulo radix into the LSBs of each pixel. No attacks on this RDH technique will be able to decode the hidden content in the marked image. The proposed scheme delivers a prominent visual quality despite high embedding capacity. Experimental tests carried out on over 100 natural image data sets and medical images show an improvement in results compared to the existing schemes. Since the algorithm is based on the variable radix number system, it is more resistant to most of the steganographic attacks. The results were compared with a higher embedding capacity of up to 1.5 bpp reversible schemes for parameters like Peak Signal-to-Noise Ratio (PSNR), Embedding Capacity (EC) and Structural Similarity Index Metric (SSIM).

Databases are an important part of today’s applications where large amounts of data need to be stored, processed, and accessed quickly. One of the important criteria when choosing to use a database technology is its data processing performance. In this paper, some methods for optimizing the database structure and queries were applied on two popular open-source database management systems: MySQL as a relational DBMS, and document-based MySQL as a non-relational DBMS. The main objective of this paper was to conduct a comparative analysis of the impact that the proposed optimization methods have on each specific DBMS when carrying out CRUD (CREATE, READ, UPDATE, DELETE) requests. To perform the analysis and performance evaluation of CRUD operations for different amounts of data, a case study testing architecture based on Java was developed and used to show how the databases’ proposed optimization methods can influence the performance of the application, and to highlight the differences in response time and complexity. The results obtained show the degree to which the proposed optimization methods contributed to the application’s performance improvement in the case of both databases; based on these, a detailed analysis and several conclusions are presented to support a decision for choosing a specific approach.

Machine Learning method is increasingly used for data analysis, where large databases has to be analyzed. One of the areas where solutions include the use of machine learning is the field of medical prediction, used to observe the probability of a person may suffer from a disease in the future. One of the domains of medicine prediction in which machine learning solutions are used is predictions in the case of diabetes. Diabetes is a disease that is increasingly present in today's society. This paper presents a comparison among the results experimentally obtained, using three machine learning algorithms in the prediction of diabetes. The three considered algorithms are support vector machine, Naive Bayes, and random forest. The aim of this paper is to analyze the performance of the algorithms considering different metrics in order to compare different techniques to obtain better accuracy. We found that support vector machine and random forest obtained an accuracy of over 80%.