Search Results (3,160)

We present GANimate, a lightweight method for animating talking faces that leverages recent advances in latent-space manipulation of Generative Adversarial Networks (GANs). Unlike existing approaches based on computationally intensive diffusion models, transformers, or complex 3DMM representations, which are impractical for mobile and other low-resource edge devices due to high memory and compute demands, GANimate is designed for efficient operation on low-memory, low-compute edge devices. The model operates on 2D lip landmarks extracted from standard mobile vision-sensor inputs and requires no pre-training, making it easily integrable with any lip-landmark generator. Through an optimization process in the GAN feature latent space, these landmarks act as geometric constraints to animate a static portrait, producing realistic and expressive lip movements. To maintain stability and visual coherence across frames, we employ a Kalman filter to detect and track lip landmarks during video synthesis, enabling adaptive refinement and improved temporal consistency. The result is a compact and modular framework that bridges the gap between performance and accessibility in talking face synthesis, delivering high-quality and stable animations with minimal computational overhead. GANimate represents an important step toward lifelike, real-time avatars suitable for sensor-enabled and mobile human–computer interaction. Full article

Traditional ergonomic considerations in human–machine and human–swarm systems have primarily relied on static diagnostic snapshots, which often fail to capture the temporal accumulation and non-linear dissipation of musculoskeletal fatigue. As Industry 5.0 transitions toward immersive, human-centric cyber-physical systems, redefining ergonomic load as an endogenous state variable allows for real-time control of musculoskeletal integrity. This work proposes the Dynamic Integrity Governor (DIG) framework, which treats ergonomic load as a normalized, dimensionless state variable $\xi \left(t\right)$ that evolves according to a stochastic proxy of recursive Newton–Euler dynamics. Leveraging a machine-perception-aware Adaptive Event-Triggered Mechanism (AETM) and the Multi-modal Flamingo Search Algorithm (MMFSA), we develop a decentralized architecture that redistributes ergonomic demands in real-time. The framework utilizes a 7-DOF kinematic model and Control Barrier Functions (CBF) to maintain human–swarm interaction within safe biomechanical boundaries, effectively filtering stochastic sensor noise through Girard-based stability buffers. Computational validation via N = 1000 Monte Carlo runs demonstrates that the proposed strategy achieves a 79.97% reduction in control updates (SD = 0.19%; p < 0.0001; Cohen’s d = 2.41), ensuring a positive minimum inter-event time (MIET) to prevent the Zeno phenomenon and supporting carbon-aware AI operations. The integration of variable prediction horizons yields an 80.69% improvement in solving time, while ensuring a minimal computational footprint suitable for real-time edge deployment. The identification of optimal postural niches maintains peak ergonomic load at 41.42%, representing a significant safety margin relative to the integrity barrier. While validated against a 50th percentile male profile, the DIG framework establishes a modular foundation for personalized ergonomic governors in inclusive Industry 5.0 applications. Full article

Background/Objectives: Advances in molecular biology, genetics, and microbiome research have significantly expanded our understanding of probiotic microorganisms and their interactions with human health, stimulating the development of both traditional and next-generation probiotic products. Although probiotics are widely used and generally considered safe for healthy individuals, accumulating evidence indicates that their safety profile varies significantly depending on the strain, dose, host, and context, with rare but clinically significant adverse events reported in vulnerable populations. Methods: This review summarizes current knowledge on the efficacy and safety of probiotics, analyzes limitations in clinical safety reporting, and compares regulatory frameworks governing the use of probiotics as dietary supplements, medicinal products, and live biotherapeutics. Particular attention is given to new genomic and computational approaches to safety assessment. Conclusions: Overall, the review emphasizes the need for coordinated regulation, rigorous clinical evidence, and integrated, modern safety assessment strategies to support the responsible expansion of probiotic use. Full article

Objectives: Global infertility rates are rapidly increasing. Assisted reproductive technologies combined with artificial intelligence are the next hope for overcoming infertility. In vitro fertilization (IVF) is gaining popularity owing to its increasing success rates. The success rate of IVF essentially depends on the assessment and inspection of blastocysts. Blastocysts can be segmented into several important compartments, and advanced and precise assessment of these compartments is strongly associated with successful pregnancies. However, currently, embryologists must manually analyze blastocysts, which is a time-consuming, subjective, and error-prone process. Several AI-based techniques, including segmentation, have been recently proposed to fill this gap. However, most existing methods rely only on raw grayscale intensity and do not perform well under challenging blastocyst image conditions, such as low contrast, similarity in textures, shape variability, and class imbalance. Methods: To overcome this limitation, we developed a novel and lightweight architecture, the microscopic multimodal interaction segmentation network (MiMics-Net), to accurately segment blastocyst components. MiMics-Net employs a multimodal blastocyst stem to decompose and process each frame into three modalities (photometric intensity, local textures, and directional orientation), followed by feature fusion to enhance segmentation performance. Moreover, MiMic dual-path grouped blocks have been designed, in which parallel-grouped convolutional paths are fused through point-wise convolutional layers to increase diverse learning. A lightweight refinement decoder is employed to refine and restore the spatial features while maintaining computational efficiency. Finally, semantic skip pathways are induced to transfer low- and mid-level spatial features after passing through the grouped and point-wise convolutional layers. Results/Conclusions: MiMics-Net was evaluated using a publicly available human blastocyst dataset and achieved a Jaccard index score of 87.9% while requiring only 0.65 million trainable parameters. Full article

In massively multiplayer online games (MMOGs), complex social environments exist in which cooperation is central not only to playing the game but also to experiencing it as an individual player. The growth of multiplayer games that emphasise cooperative activities in computer-based environments has sparked academic interest in collaboration and its role in the field, engaging scholars from domains such as human–computer interaction and digital entertainment. This paper presents a systematic literature review (SLR) and bibliometric analysis of 70 peer-reviewed journal papers published between 2015 and 2024. This data is derived from the Web of Science and Scopus databases. This literature review contributes to the understanding of collaborative factors in MMOGs, which include task interdependence, communication, trust, leadership, and player behaviour. The review is in the field using bibliometrics. To present the findings, we construct an input–process–output (IPO) model that links game features (inputs) and interaction dynamics (processes) to team performance and player experience (outputs) in MMOGs. This review maps the field’s dominant factors (task interdependence, communication, trust, leadership, and player behaviour), pinpoints methodological priorities, and sets a concrete agenda for future research on team collaboration in MMOGs. Full article

Previous studies have emphasized the role of task difficulty in decision performance while relatively neglecting the decision maker’s subjective initiative and intrinsic learning process during task execution. This study manipulated the rule hierarchy factor, which reflects external task difficulty, and the block factor, which reflects the accumulation of intrinsic learning, and used analysis of variance (ANOVA), the drift-diffusion model (DDM), and random forest algorithms to systematically examine how task difficulty and learning jointly influence decision behavior and its underlying mechanisms. A total of 40 participants were recruited, and after strict exclusion criteria were applied, 34 valid datasets were included in the final analysis. The results showed that although rule hierarchy had a significant impact on decision performance in the early stage of the task (the first two blocks), this effect gradually diminished as task repetitions increased. Furthermore, the results revealed a clear dissociation in predictive mechanisms: intrinsic cognitive factors (specifically, evidence accumulation efficiency and decision bias) were the primary predictors of decision accuracy, whereas external task difficulty (rule hierarchy) acted as the dominant predictor for decision speed (reaction time). These findings provide a new perspective for understanding the dynamic relationship between external task demands and intrinsic learning processes, highlighting the necessity of distinguishing between accuracy and speed metrics in personalized education, training, and human–computer interaction design. Full article

Background/Objectives: The tumor suppressor gene TP53 is one of the most frequently mutated genes in human cancers, with alterations predominantly affecting its DNA-binding domain (DBD). However, the mutational landscape and functional consequences of TP53 variants remain poorly characterized in African populations. This study aimed to characterize mutations in exons 5–6 of TP53 in oral cavity cancer (OCC), prostate cancer (PC), and breast cancer (BC) in a Senegalese population, and to assess their structural effects, functional consequences, and impact on protein–protein interactions with BCL-2. Methods: Seventy-eight archived tumor DNA samples from Senegalese patients with OCC, PC, and BC were analyzed. Variants were annotated using COSMIC and dbSNP databases. Functional impact was evaluated with PolyPhen-2. Structural stability changes ($\Delta \Delta G$) were predicted using FoldX, conformational dynamics ($\Delta \Delta S_{vib}$) were assessed with ENCoM, and effects on the p53–BCL-2 interaction were analyzed using DDMut-PPI. Statistical analyses were also performed. Results: BC exhibited the highest TP53 mutation frequency, whereas OCC showed greater mutational diversity. Exon-level analysis revealed a significant enrichment of exon 6 mutations in BC. Structural analyses indicated that exon 5 mutations across all cancers and mutations in OCC were predominantly destabilizing and associated with loss-of-function effects. In contrast, recurrent exon 6 mutations in PC and BC, particularly V217L and V218M, were predicted to stabilize the p53 structure. Conformational dynamics differences between exons were significant only in PC. All analyzed mutations were predicted to stabilize the p53–BCL-2 interaction. Conclusions: This integrative in silico study identified cancer and exon-specific TP53 mutation patterns in a Senegalese population, highlighting exon 6 as a context-dependent hotspot with potential oncogenic implication in PC and BC. Despite its computational nature, the study provides valuable insights that merit further investigation. Full article

Current generative AI systems primarily utilize a prompt–response interaction model that restricts user intervention during the creative process. This lack of granular control creates a significant disconnect between user intent and machine output, which we define as the “Agency Gap”. This paper introduces the Agency-First Framework (AFF), which combines cognitive engineering and co-active design approaches to formally define human-AI collaboration. This is operationalized through the development of ten Generative AI Agency (GAIA) Heuristics, a systematic method for evaluating agency-centric interactions within stochastic generative settings. By translating the theoretical layers of the AFF into measurable criteria, the GAIA heuristics provide the necessary instrument for the empirical auditing of existing systems and the guidance of agency-centric redesigns. Unlike existing assistive AI guidelines that focus on output-level usability, the AFF establishes agency as a first-class design construct, enabling mid-process intervention and the steering of the model’s latent reasoning trajectory. Validation of the AFF was conducted through a two-tiered empirical evaluation: (1) an expert heuristic audit of state-of-the-art platforms, such as ChatGPT-o1 and Midjourney v6, which achieved high inter-rater reliability, and (2) a controlled redesign study. The latter demonstrated that agency-centric interfaces significantly enhance the Sense of Agency and Intent Alignment Accuracy compared to baseline prompt-response models, even when introducing a deliberate increase in task completion time—a phenomenon we describe as “productive friction” or an intentional interaction slowdown designed to prioritize cognitive engagement and user control over raw speed. Overall, the findings suggest that the restoration of meaningful user agency requires a shift from “seamless” system efficiency towards “productive friction”, where controllability and transparency within the generative process are prioritized. The major contribution of this work is the provision of a scalable, empirically validated framework and set of heuristics that equip designers to move beyond prompt-centric interaction, establishing a methodological foundation for agency-preserving generative AI systems. Full article

The MERS-CoV (Middle East respiratory syndrome coronavirus) is a zoonotic virus with a high mortality rate and a lack of antiviral drugs, underscoring the need for effective therapeutic methods. Viral entry depends on interactions between viral surface proteins and human receptors, with Dipeptidyl Peptidase-4 (DPP4), a transmembrane glycoprotein, acting as the receptor for MERS-CoV. We employed Molecular Dynamics (MD) Simulations to identify critical interface residues under a high-performance computing (HPC) workflow for accelerated results. Target residue pairs were identified through analysis of salt bridge and hydrogen bond occupancy. The stability of these residues was confirmed through three independent MD Simulations at human body temperature and constant pressure. Additionally, binding affinity predictions were calculated to determine the interaction strength between the virus and human receptors. Applying the scientific threshold criteria, we narrowed our results to seven key interaction pairs; two of the identified pairs (Asp510-Arg317, and Arg511-Asp393) are consistent with findings published in previous research studies, and five novel interactions are proposed for future experimental studies with our active collaborators in Pharmacology. The results provide a molecular basis for targeted mutation-based experiments and support the rational design of structure-based inhibitors aimed at disrupting the MERS-CoV-DPP4 complex, thereby facilitating the translation of computational findings into antiviral drug discovery. Full article

This project presents the development of a Retrieval-Augmented Generation (RAG) system applied to the customization of a Non-Playable Character (NPC), designed as the Non-Playable Character (NPC) of the President of the IDIS Foundation Città della Scienza (City of Science). The NPC acts as both a virtual guide and institutional ambassador within the science center, providing multilingual, interactive, and accessible communication for a broad international audience. Through the integration of generative models with a curated, validated knowledge base, the RAG system enables the NPC to provide accurate, context-sensitive, and up-to-date responses to user queries. Developed by the Teaching Learning Centre for Education and Inclusive Technologies ‘Elisa Frauenfelder’ at the University of Salerno, the system supports the museum’s educational mission by enhancing science communication and fostering inclusive digital engagement. The Non-Playable Character (NPC) features realistic facial animation, movement, and voice synthesis, creating a digital twin capable of simulating human-like interaction. This initiative exemplifies an innovative application of artificial intelligence for an inclusive and equitable quality education and contributes to the development of engaging, accessible, and personalized learning environments. Full article

Temporal action detection (TAD) aims to localize and recognize action instances in untrimmed videos, and serves as a key component in practical intelligent electronic systems such as smart video surveillance and real-time human–machine interaction. In these scenarios, accurate temporal localization is essential for reliable event understanding and downstream decision-making in edge computing and real-time streaming scenarios. To handle long video durations and diverse action dynamics, existing methods typically rely on hierarchical temporal feature integration to capture multi-scale contextual information. However, such integration often leads to intra-segment inconsistency and boundary ambiguity, as indiscriminate temporal smoothing across scales degrades segment coherence and blurs critical boundary cues. In this work, we propose FreqAct, a multi-frequency feature fusion framework that explicitly models complementary low-frequency and high-frequency temporal components within hierarchical representations. Specifically, low-frequency modulation suppresses undesired temporal fluctuations to stabilize segment-level representations, while high-frequency enhancement preserves boundary-sensitive cues essential for precise localization. Furthermore, we introduce a boundary-aware regression loss to emphasize learning at action boundaries and an intra-segment consistency regularization to encourage coherent predictions within each action instance. Extensive experiments on THUMOS14 and ActivityNet1.3 demonstrate that FreqAct outperforms state-of-the-art TAD methods, further highlighting its effectiveness and practical potential for real-world electronics applications. Full article

Self-organizing systems arise in complex biomechanical structures, human locomotion, and neural control hierarchies, yet quantitative methods for describing order formation and loss of stability remain limited. This study develops a mathematical framework for analyzing self-organization using entropy-based measures, indicators of chaotic dynamics, and network-theoretic structure. The approach (the LET framework) combines Lyapunov exponents with entropy families and graph metrics (algebraic connectivity, Load-Path Heterogeneity Index) to: (i) examine transitions between ordered and disordered states, (ii) assess sensitivity to perturbations, and (iii) characterize structural coherence in evolving cervical spine kinematics. Analytical models and computational validations are presented for cervical stability and post-operative Adjacent Segment Disease (ASD) using the Branney–Breen dataset. The findings indicate that entropy and chaos measures identify regime shifts and the emergence of a “stability corridor” more clearly than task-oriented indices, and provide finer resolution of dynamical variability within self-organizing processes. Network metrics complement these results by linking local segmental interactions to global structural fragility transfer. The study shows that entropy, chaos indicators, and network structure together form a consistent basis for describing self-organization in biomechanical systems, enabling quantitative comparison of dynamical regimes and improved interpretation of emergent pathological behavior. The approach utilizes a hybrid kinematic surrogate model to resolve passive and active components, bypassing direct force measurements by employing viscoelastic mechanotransduction principles. Full article

Modern interaction with smart devices is hindered by the “Midas Touch” problem, where sensors frequently misinterpret incidental physical movements as intentional commands due to a lack of human context. This research addresses this conflict by introducing the Multimodal Haptic Informatics (MHI) index within a novel Scene–Action–Trigger (SAT) framework. The goal is to contextualize mechanical movements as human intent by integrating physical, spatial, and cognitive data locally at the edge. The methodology employs an “Action-as-primary indexing” mechanism where the Action channel (IMU) serves as a temporal anchor t, triggering high-resolution Scene (computer vision) and Trigger (audio) processing only during critical haptic events. Validated through a complex origami crane task generating 29,408 data frames, the framework utilizes a three-stage informatics derivation process: single-modal scoring, score weighting, and hand state mapping. Results demonstrate that applying an adaptive “Speedometer” logic successfully reclassifies the “Transitional State”. While this state constitutes over half of the behavioral dataset (54.76% on average), it is effectively disambiguated into meaningful intent using a self-trained local Large Language Model (LLM) for semantic verification. Furthermore, the event-driven sampling of 93 keyframes reduces the processing overhead by 99.68% compared to linear annotation. This study contributes a low-latency, privacy-preserving “Protocol of Assent” that maintains user agency by providing intelligent system suggestions based on confirmed haptic intensity. Full article

The integration of Generative Artificial Intelligence (GenAI) into higher education represents a paradigm shift from static skill acquisition to dynamic, human–AI collaboration. However, the psychological mechanisms governing students’ adoption—specifically the interplay between pedagogical novelty, ethical awareness, and habit formation—remain underexplored. To address this, this study develops and implements a dynamic practical curriculum incorporating AI and ethical awareness, aiming to foster responsible behavioral patterns in computer programming education. Employing a quasi-experimental design, we implemented a 16-week dual-track instructional intervention (incorporating AI-integrated pedagogy and ethical scaffolding) for 148 computer science students. Structural Equation Modeling (SEM) was applied to test an extended UTAUT2 framework. The findings reveal three critical theoretical insights that redefine GenAI adoption: (1) The eclipse of utility: contrary to established models, traditional utilitarian drivers of performance expectancy (β = 0.076, p = 0.39) and effort expectancy (β = 0.125, p = 0.13) yielded non-significant effects on behavioral intention. This suggests that for digital natives, algorithmic efficiency has devolved into a baseline hygiene factor, losing its motivational power. (2) The dominance of pedagogical novelty: hedonic motivation emerged as the paramount predictor of both habit (β = 0.457, p < 0.001) and behavioral intention (β = 0.336, p = 0.001). This confirms that adoption is driven by the situational interest and interactional novelty inherent in the human–AI partnership. (3) The cognitive brake mechanism: ethical awareness exhibited a divergent regulatory role. While it significantly legitimized conscious behavioral intention (β = 0.166, p = 0.011), it showed a non-significant, negative association with habit (β = −0.032, p = 0.653). This demonstrates that ethical reasoning functions as a cognitive brake (system 2) and actively disrupts the formation of mindless, automated dependency (system 1). These results provide empirical evidence for a dual regulation model of AI adoption and suggest that sustainable education requires leveraging pedagogical novelty to drive engagement while utilizing ethical awareness to prevent blind habituation. Full article

The doctor–patient relationship is a central factor in healthcare delivery. Artificial Intelligence (AI) represents an emerging technological frontier whose implications remain to be fully clarified. Evidence-based studies provide reliable analyses of effects and offer a deeper understanding of both limits and benefits. This narrative review aimed to explore the role of AI in modern clinical practice, with particular reference to its effects on the doctor–patient relationship. Scopus and Web of Science databases were searched between 1 and 10 December 2025 to identify suitable studies. Inclusion criteria comprised English-language articles published in the last 10 years, with a direct focus on the doctor–patient relationship and exclusively employing empirical research designs. A total of 21 studies published between 2021 and 2025 were identified as eligible. The most common AI applications were conceptual systems discussed at a perceptual level (thirteen studies), followed by simulated AI decision-making scenarios (two studies). Implemented AI applications were less frequent and mainly included AI-based clinical decision support systems, administrative and documentation-focused tools, and a small number of conversational or relational AI applications (six studies in total). These studies focused on patients, healthcare professionals, and medical students preparing for future clinical roles. Results highlighted generally positive patient attitudes toward AI, often mediated by educational level, technological familiarity, and risk awareness. Among healthcare professionals, positive attitudes also emerged, although concerns regarding epistemic and professional values were noted. Greater involvement of clinicians in its development was consistently recommended. Findings from academic samples aligned with those of patients and clinicians, showing that integrating AI with traditional clinical practices was consistently preferred. Empathy, compassion, effective communication, accuracy, ethics, and trust were highlighted as fundamental values essential for mitigating risks. These elements are fundamental to the effective implementation of technologies aimed at improving clinical practice, while an integrative perspective is needed to safeguard the doctor–patient relationship. Overall, the use of AI in medical practice emerged as promising. Further studies should strengthen the empirical basis of the field to support an evidence-based approach to AI integration in healthcare. Full article