Search Results (852)

Media platforms provide an effective way to gauge public perceptions, especially during mass disruption events. This research explores public responses to the 2023 Ohio train derailment event through Twitter, currently known as X, and Google Trends. It aims to unveil public sentiments and attitudes by employing sentiment analysis using the Valence Aware Dictionary and Sentiment Reasoner (VADER) and topic modeling using Latent Dirichlet Allocation (LDA) on geotagged tweets across three phases of the event: impact and immediate response, investigation, and recovery. Additionally, the Self-Organizing Map (SOM) model is employed to conduct time-series clustering analysis of Google search patterns, offering a deeper understanding into the event’s spatial and temporal impact on society. The results reveal that public perceptions related to pollution in communities exhibited an inverted U-shaped curve during the initial two phases on both the Twitter and Google Search platforms. However, in the third phase, the trends diverged. While public awareness declined on Google Search, it experienced an uptick on Twitter, a shift that can be attributed to governmental responses. Furthermore, the topics of Twitter discussions underwent a transition across three phases, changing from a focus on the causes of fires and evacuation strategies in Phase 1, to river pollution and trusteeship issues in Phase 2, and finally converging on government actions and community safety in Phase 3. Overall, this study advances a multi-platform and multi-method framework to uncover the spatiotemporal dynamics of public perception during disasters, offering actionable insights for real-time, region-specific crisis management. Full article

As a foundation of large language models, fine-tuning drives rapid progress, broad applicability, and profound impacts on human–AI collaboration, surpassing earlier technological advancements. This paper provides a comprehensive overview of large language model (LLM) fine-tuning by integrating hermeneutic theories of human comprehension, with a focus on the essential cognitive conditions that underpin this process. Drawing on Gadamer’s concepts of Vorverständnis, Distanciation, and the Hermeneutic Circle, the paper explores how LLM fine-tuning evolves from initial learning to deeper comprehension, ultimately advancing toward self-awareness. It examines the core principles, development, and applications of fine-tuning techniques, emphasizing its growing significance across diverse field and industries. The paper introduces a new term, “Tutorial Fine-Tuning (TFT)”, which annotates a process of intensive tuition given by a “tutor” to a small number of “students”, to define the latest round of LLM fine-tuning advancements. By addressing key challenges associated with fine-tuning, including ensuring adaptability, precision, credibility and reliability, this paper explores potential future directions for the co-evolution of humans and AI. By bridging theoretical perspectives with practical implications, this work provides valuable insights into the ongoing development of LLMs, emphasizing their potential to achieve higher levels of cognitive and operational intelligence. Full article

The interaction of estrogen receptor alpha (ERα) with various metabolites—both endogenous and exogenous, such as those present in food products, as well as gut microbiota-derived metabolites—plays a critical role in modulating the hormonal balance in the human body. In this study, we evaluated a suite of 27 machine learning models and, following systematic optimization and rigorous performance comparison, identified linear discriminant analysis (LDA) as the most effective predictive approach. A meticulously curated dataset comprising 75 molecular descriptors derived from compounds with known ERα activity was assembled, enabling the model to achieve an accuracy of 89.4% and an F1 score of 0.93, thereby demonstrating high predictive efficacy. Feature importance analysis revealed that both topological and physicochemical descriptors—most notably FractionCSP3 and AromaticProportion—play pivotal roles in the potential binding to ERα. Subsequently, the model was applied to chemicals commonly encountered in food products, such as indole and various phenolic compounds, indicating that approximately 70% of these substances exhibit activity toward ERα. Moreover, our findings suggest that food processing conditions, including fermentation, thermal treatment, and storage parameters, can significantly influence the formation of these active metabolites. These results underscore the promising potential of integrating predictive modeling into food technology and highlight the need for further experimental validation and model refinement to support innovative strategies for developing healthier and more sustainable food products. Full article

Federated learning is the on-device, collaborative training of a global model that can be utilized to support the privacy preservation of participants’ local data. In federated learning, there are challenges to model training regarding privacy preservation, security, resilience, and integrity. For example, a malicious server can indirectly obtain sensitive information through shared gradients. On the other hand, the correctness of the global model can be corrupted through poisoning attacks from malicious clients using carefully manipulated updates. Many related works on secure aggregation and poisoning attack detection have been proposed and applied in various scenarios to address these two issues. Nevertheless, existing works are based on the trust confidence that the server will return correctly aggregated results to the participants. However, a malicious server may return false aggregated results to participants. It is still an open problem to simultaneously preserve users’ privacy and defend against poisoning attacks while enabling participants to verify the correctness of aggregated results from the server. In this paper, we propose a privacy-preserving and poisoning attack-resilient federated learning framework that supports the verification of aggregated results from the server. Specifically, we design a zero-trust dual-server architectural framework instead of a traditional single-server scheme based on trust. We exploit additive secret sharing to eliminate the single point of exposure of the training data and implement a weight selection and filtering strategy to enhance robustness to poisoning attacks while supporting the verification of aggregated results from the servers. Theoretical analysis and extensive experiments conducted on real-world data demonstrate the practicability of our proposed framework. Full article

The $\delta$-GLMB filter is an analytic solution to the multi-target Bayes recursion used in multi-target tracking. It extends the Generalised Labelled Multi-Bernoulli (GLMB) framework by providing an efficient and scalable implementation while preserving track identities, making it a widely used approach in the field. Theoretically, the $\delta$-GLMB filter handles uncertainties in measurements in its filtering procedure. However, in practice, degeneration of the measurement quality affects the performance of this filter. In this paper, we discuss the effects of increasing measurement uncertainty on the $\delta$-GLMB filter and also propose two heuristic methods to improve the performance of the filter in such conditions. The base idea of the proposed methods is to utilise the information stored in the history of the filtering procedure, which can be used to decrease the measurement uncertainty effects on the filter. Since GLMB filters have shown good results in the field of multi-target tracking, an uncertainty-immune $\delta$-GLMB can serve as a strong tool in this area. In this study, the results indicate that the proposed heuristic ideas can improve the performance of filtering in the presence of uncertain observations. Experimental evaluations demonstrate that the proposed methods enhance track continuity and robustness, particularly in scenarios with low detection rates and high clutter, while maintaining computational feasibility. Full article

This paper introduces a model that automates the diagnosis of a patient’s condition, reducing reliance on highly trained professionals, particularly in resource-constrained settings. To ensure data consistency, the dataset was preprocessed for uniformity in size, format, and color channels. Image quality was further enhanced using histogram equalization to improve the dynamic range. Lung regions were isolated using segmentation techniques, which also eliminated extraneous areas from the images. A modified segmentation-based cropping technique was employed to define an optimal cropping rectangle. Feature extraction was performed using persistent homology, deep learning, and hybrid methodologies. Persistent homology captured topological features across multiple scales, while the deep learning model leveraged convolutional transition equivariance, input-adaptive weighting, and the global receptive field provided by Vision Transformers. By integrating features from both methods, the classification model effectively predicted severity levels (mild, moderate, severe). The segmentation-based cropping method showed a modest improvement, achieving 80% accuracy, while stand-alone persistent homology features reached 66% accuracy. Notably, the hybrid model outperformed existing approaches, including SVM, ResNet50, and VGG16, achieving an accuracy of 82%. Full article

Traditional image denoising algorithms often struggle with real-world complexities such as spatially correlated noise, varying illumination conditions, sensor-specific noise patterns, motion blur, and structural distortions. This paper presents an enhanced residual denoising network, R-REDNet, which stands for Reinforced Residual Encoder–Decoder Network. The proposed architecture incorporates deeper convolutional layers in the encoder and replaces additive skip connections with averaging operations to improve feature extraction and noise suppression. Additionally, the method leverages an iterative refinement approach, further enhancing its denoising performance. Experiments conducted on two real-world noisy image datasets demonstrate that R-REDNet outperforms current state-of-the-art approaches. Specifically, it attained a peak signal-to-noise ratio of 44.01 dB and a structural similarity index of 0.9931 on Dataset 1, and it obtained a peak signal-to-noise ratio of 46.15 dB with a structural similarity index of 0.9955 on Dataset 2. These findings confirm the efficiency of our method in delivering high-quality image restoration while preserving fine details. Full article

Detecting mentions of green waste practices on social networks is a crucial tool for environmental monitoring and sustainability analytics. Social media serve as a valuable source of ecological information, enabling researchers to track trends, assess public engagement, and predict the spread of sustainable behaviors. Automatic extraction of mentions of green waste practices facilitates large-scale analysis, but the uneven distribution of such mentions presents a challenge for effective detection. To address this, data augmentation plays a key role in balancing class distribution in green practice detection tasks. In this study, we compared existing data augmentation techniques based on the paraphrasing of original texts. We evaluated the effectiveness of additional explanations in prompts, the Chain-of-Thought prompting, synonym substitution, and text expansion. Experiments were conducted on the GreenRu dataset, which focuses on detecting mentions of green waste practices in Russian social media. Our results, obtained using two instruction-based large language models, demonstrated the effectiveness of the Chain-of-Thought prompting for text augmentation. These findings contribute to advancing sustainability analytics by improving automated detection and analysis of environmental discussions. Furthermore, the results of this study can be applied to other tasks that require augmentation of text data in the context of ecological research and beyond. Full article

In recent years, the healthcare market has grown very fast and is dealing with a huge increase in data. Healthcare applications are time-sensitive and need quick responses with fewer delays. Fog Computing (FC) was introduced to achieve this aim. It can be applied in various application areas like healthcare, smart and intelligent environments, etc. In healthcare applications, some tasks are considered critical and need to be processed first; other tasks are time-sensitive and need to be processed before their deadline. In this paper, we have proposed a Task Classification algorithm based on Deadline and Criticality (TCDC) for serving healthcare applications in a fog environment. It depends on classifying tasks based on the critical level to process critical tasks first and considers the deadline of the task, which is an essential parameter to consider in real-time applications. The performance of TCDC was compared with some of the literature. The simulation results showed that the proposed algorithm can improve the overall performance in terms of some QoS parameters like makespan with an improved ratio from 60% to 70%, resource utilization, etc. Full article

In this paper, we propose an improved version of the Pyramidal Predictive Network (PPNV2), a theoretical framework inspired by predictive coding, which addresses the limitations of its predecessor (PPNV1) in the task of future perception prediction. While PPNV1 employed a temporal pyramid architecture and demonstrated promising results, its innate signal processing led to aliasing in the prediction, restricting its application in robotic navigation. We analyze the signal dissemination and characteristic artifacts of PPNV1 and introduce architectural enhancements and training strategies to mitigate these issues. The improved architecture focuses on optimizing information dissemination and reducing aliasing in neural networks. We redesign the downsampling and upsampling components to enable the network to construct images more effectively from low-frequency-input Fourier features, replacing the simple concatenation of different inputs in the previous version. Furthermore, we refine the training strategies to alleviate input inconsistency during training and testing phases. The enhanced model exhibits increased interpretability, stronger prediction accuracy, and improved quality of predictions. The proposed PPNV2 offers a more robust and efficient approach to future video-frame prediction, overcoming the limitations of its predecessor and expanding its potential applications in various robotic domains, including pedestrian prediction, vehicle prediction, and navigation. Full article

This study explores factors influencing positive experiences with generative AI (GenAI) in a learning game design context. Using a sample of 26 master’s-level students in a course on AI’s societal aspects, this study examines the impact of (1) prior knowledge and attitudes toward technology and learning, and (2) personal value orientations. Results indicated that both students’ self-transcendent goals and desire for material benefits have positive correlations with collaborative, cognitive, and affective outcomes. However, self-transcendent goals are a stronger predictor, as determined by stepwise regression analysis. Attitudes toward technology were positively associated with cognitive and affective outcomes during the first week, though this association did not persist into the second week. Most other attitudinal variables were not associated with collaborative or cognitive outcomes but were linked to negative affect. These findings suggest that students’ personal values correlate more strongly with the collaborative, cognitive, and affective aspects of using GenAI for educational game design than their attitudinal attributes. This result may indicate that the design experience neutralizes the effect of earlier attitudes towards technology, with major influences deriving from personal value orientations. If these findings are borne out, this study has implications for the utility of current educational efforts to change students’ attitudes towards technology, especially those that encourage more women to study STEM topics. Thus, it may be that, rather than pro-technology instruction, a focus on value orientations would be a more effective way to encourage diverse students to participate in STEM programs. Full article

With the rapid increase in the amount of big data, traditional software tools are facing complexity in tackling big data, which is a huge concern in the research industry. In addition, the management and processing of big data have become more difficult, thus increasing security threats. Various fields encountered issues in fully making use of these large-scale data with supported decision-making. Data mining methods have been tremendously improved to identify patterns for sorting a larger set of data. MapReduce models provide greater advantages for in-depth data evaluation and can be compatible with various applications. This survey analyses the various map-reducing models utilized for big data processing, the techniques harnessed in the reviewed literature, and the challenges. Furthermore, this survey reviews the major advancements of diverse types of map-reduce models, namely Hadoop, Hive, Pig, MongoDB, Spark, and Cassandra. Besides the reliable map-reducing approaches, this survey also examined various metrics utilized for computing the performance of big data processing among the applications. More specifically, this review summarizes the background of MapReduce and its terminologies, types, different techniques, and applications to advance the MapReduce framework for big data processing. This study provides good insights for conducting more experiments in the field of processing and managing big data. Full article

Quantum key distribution (QKD), a key application of quantum information technology and “one-time pad” (OTP) encryption, enables secure key exchange with information-theoretic security, meaning its security is grounded in the laws of physics rather than computational assumptions. However, in QKD networks, achieving long-distance communication often requires trusted relays to mitigate channel losses. This reliance introduces significant challenges, including vulnerabilities to compromised relays and the high costs of infrastructure, which hinder widespread deployment. To address these limitations, we propose a zero-trust spatiotemporal diversification framework for multipath–multi-key distribution. The proposed approach enhances the security of end-to-end key distribution by dynamically shuffling key exchange routes, enabling secure multipath key distribution. Furthermore, it incorporates a dynamic adaptive path recovery mechanism that leverages a recursive penalty model to identify and exclude suspicious or compromised relay nodes. To validate this framework, we conducted extensive simulations and compared its performance against established multipath QKD methods. The results demonstrate that the proposed approach achieves a 97.22% lower attack success rate with 20% attacker pervasiveness and a 91.42% reduction in the attack success rate for single key transmission. The total security percentage improves by 35% under 20% attacker pervasiveness, and security enhancement reaches 79.6% when increasing QKD pairs. Additionally, the proposed scheme exhibits an 86.04% improvement in defense against interception and nearly doubles the key distribution success rate compared to traditional methods. The results demonstrate that the proposed approach significantly improves both security robustness and efficiency, underscoring its potential to advance the practical deployment of QKD networks. Full article

This study explores explicit and implicit knowledge in large-scale linguistic data and digital footprints from social networks. This research aims to develop and test algorithms for analyzing both explicit and implicit information in user-generated content and digital interactions. A dataset of social media discussions on avian influenza in Moscow (RF) was collected and analyzed (tokens: 1,316,387; engagement: 108,430; audience: 39,454,014), with data collection conducted from 1 March 2023, 00:00 to 31 May 2023, 23:59. This study employs Brand Analytics, TextAnalyst 2.32, ChatGPT o1, o1-mini, AutoMap, and Tableau as analytical tools. The findings highlight the advantages and limitations of explicit and implicit information analysis for social media data interpretation. Explicit knowledge analysis is more predictable and suitable for tasks requiring quantitative assessments or classification of explicit data, while implicit knowledge analysis complements it by enabling a deeper understanding of subtle emotional and contextual nuances, particularly relevant for public opinion research, social well-being assessment, and predictive analytics. While explicit knowledge analysis provides structured insights, it may overlook hidden biases, whereas implicit knowledge analysis reveals underlying issues but requires complex interpretation. The research results emphasize the importance of integrating various scientific paradigms and artificial intelligence technologies, particularly large language models (LLMs), in the analysis of social networks. Full article

Sports analytics is a fast-evolving domain using advanced data science methods to find useful insights. This study explores the way NBA player performance metrics evolve from quarter to quarter and affect game outcomes. Using Association Rule Mining, we identify key offensive, defensive, and overall impact metrics that influence success in both regular-season and playoff contexts. Defensive metrics become more critical in late-game situations, while offensive efficiency is paramount in the playoffs. Ball handling peaks in the second quarter, affecting early momentum, while overall impact metrics, such as Net Rating and Player Impact Estimate, consistently correlate with winning. In the collected dataset we performed preprocessing, applying advanced anomaly detection and discretization techniques. By segmenting performance into five categories—Offense, Defense, Ball Handling, Overall Impact, and Tempo—we uncovered strategic insights for teams, coaches, and analysts. Results emphasize the importance of managing player fatigue, optimizing lineups, and adjusting strategies based on quarter-specific trends. The analysis provides actionable recommendations for coaching decisions, roster management, and player evaluation. Future work can extend this approach to other leagues and incorporate additional contextual factors to refine evaluation and predictive models. Full article