Search Results (646)

Legged robots have great potential in complex environments, but achieving robust and natural locomotion remains difficult due to challenges in generating smooth gaits and resisting disturbances. This article presents a novel reinforcement learning framework that integrates a skeleton-aware graph neural network (GNN), a single-stage teacher–student architecture, a system-response model, and a Wasserstein Adversarial Motion Priors (wAMP) module. The skeleton-aware GNN enriches observations by encoding key node information and link properties, providing structured body information and better spatial awareness on irregular terrains. Unlike conventional two-stage approaches, this method jointly trains teacher and student policies to accelerate learning and improve sim-to-real transfer using hybrid advantage estimation (HAE). The system-response model further enhances robustness by predicting future observations from historical states via contrastive learning, enabling the policy to anticipate terrain variations and external disturbances. Finally, wAMP provides a more stable adversarial imitation method for fitting expert datasets of both flat ground and stair locomotion. Experiments on quadruped robots demonstrate that the proposed approach achieves more natural gaits and stronger robustness than existing baselines. Full article

Background: This paper concerns the study of forces acting on the upper ankle joint of a human in static and quasi-dynamic positions. This paper aimed to determine the pressure forces on the axis of the upper ankle joint in the position of the body tilting forward and backward, as well as in a neutral position. Methods: A model with designated centres of gravity (including and excluding the weight of the platform imitating the foot) and the point of gravity imitating the proximal insertion of the triceps surae and tibialis anterior muscles was developed for this study. The forces and the weight of the tilted object were measured using dynamometers. A method for determining the arms of gravitational forces and the angle of inclination of an object is presented. The function describing the distribution of gravitational loading along its tilting part was described. Next, all measurements and calculations were referred to the human body. Results: Measurements of muscle force, body gravity, the arms of these forces, and the angles of the object’s inclination on the axis of rotation are presented. A methodology for determining the pressure force on the human upper ankle joint axis is presented. The distribution of the value of the pressure force and its components from the maximal forward, through the vertical body position, up to the maximal backward position of the body tilt, is provided. Conclusions: The ankle joint pressure force is the vector sum of the force of gravity and the force of the muscle counteracting the body tilt. This force is the smallest in the vertical body position and increases with the body tilt. It reaches 5.23 times the weight of the tilting part of the body when the body is tilted to its maximum forward position, and 3.57 times the weight when the body tilts backward. Regardless of the direction of the body tilt, the joint pressure vector always runs through the axis of the upper ankle joint. Full article

This paper presents TVAE-SSL, a novel semi-supervised learning (SSL) paradigm that involves Tabular Variational Autoencoder (TVAE)-sampled synthetic data injection into the training process to enhance model performance under low-label data conditions in Educational Data Mining tasks. The algorithm begins with training a TVAE on the given labeled data to generate imitative synthetic samples of the underlying data distribution. These synthesized samples are treated as additional unlabeled data and combined with the original unlabeled ones in order to form an augmented training pool. A standard SSL algorithm (e.g., Self-Training) is trained using a base classifier (e.g., Random Forest) on the combined dataset. By expanding the pool of unlabeled samples with realistic synthetic data, TVAE-SSL improves training sample quantity and diversity without introducing label noise. Large-scale experiments on a variety of datasets demonstrate that TVAE-SSL can outperform baseline supervised models in the full labeled dataset in terms of accuracy, F1-score and fairness metrics. Our results demonstrate the capacity of generative augmentation to enhance the effectiveness of semi-supervised learning for tabular data. Full article

This paper presents a methodological framework for selectively optimizing computer vision models for safety-critical applications. Through systematic processes of hyperparameter tuning alongside multitask learning, we attempt to create a highly interpretable system to better assess the dangers of models intended for safety operations and intentionally select their trade-offs. Using thermographic images of a specific imitation explosive, we create a case study for the viability of humanitarian demining operations. We hope to demonstrate how this approach provides a developmental framework for creating humanitarian AI systems that optimize safety verification in real-world scenarios. By employing a comprehensive grid search across 64 model configurations to evaluate how loss function weights impact detection reliability, with particular focus on minimizing false negative rates due to their operational impact. The optimized configuration achieves a 37.5% reduction in false negatives while improving precision by 2.8%, resulting in 90% detection accuracy with 92% precision. However, to expand the generalizability of this model, we hope to call institutions to openly share their data to increase the breadth of imitation landmines and terrain data to train models from. Full article

Background: Evolutionary game theory (EGT), originating from Darwinian competition studies, offers a powerful framework for understanding complex healthcare interactions where multiple stakeholders with conflicting interests evolve strategies over time. Unlike traditional game theory, EGT accounts for bounded rationality and strategic evolution through imitation and selection. Aims and objectives: In our study, we use Synthetic Knowledge Synthesis (SKS) that integrates descriptive bibliometrics and bibliometric mapping to systematically analyze the application of EGT in healthcare. The SKS aimed to identify prolific research topics, suitable publishing venues, and productive institutions/countries for collaboration and funding. Data was harvested from the Scopus bibliographic database, encompassing 539 publications from 2000 to June 2025, Results: Production dynamics is revealing an exponential growth in scholarly output since 2019, with peak productivity in 2024. Descriptive bibliometrics showed China as the most prolific country (376 publications), followed by the United States and the United Kingdom. Key institutions are predominantly Chinese, and top journals include PLoS One and Frontiers in Public Health. Funding is primarily from Chinese entities like the National Natural Science Foundation of China. Bibliometric mapping identified five key research themes: game theory in cancer research, evolution game-based simulation of supply management, evolutionary game theory in epidemics, evolutionary games in trustworthy connected public health, and evolutionary games in collaborative governance. Conclusions: Despite EGT’s utility, significant research gaps exist in methodological robustness, data availability, contextual modelling, and interdisciplinary translation. Future research should focus on integrating machine learning, longitudinal data, and explicit ethical frameworks to enhance EGT’s practical application in adaptive, patient-centred healthcare systems. Full article

Differentiating genuine hemiplegic gait (HG) in stroke survivors from hemiplegic-like gait voluntarily imitated by healthy adults (MHG) is essential for reliable assessment and intervention planning. Treadmill-based gait data were obtained from 79 participants—39 stroke patients (HG) and 40 healthy adults—instructed to mimic HG (MHG). Forty-eight spatiotemporal and force-related variables were extracted. Random Forest, support vector machine (SVM), and logistic regression classifiers were trained with (i) the full feature set and (ii) the 10 most important features selected via Random Forest Gini importance. Performance was assessed with 5-fold stratified cross-validation and an 80/20 hold-out test, using accuracy, F1-score, and the area under the receiver operating characteristic curve (AUC). All models achieved high discrimination (AUC > 0.93). The SVM attained perfect discrimination (AUC = 1.000, test set) with the full feature set and maintained excellent accuracy (AUC = 0.983) with only the top 10 features. Temporal asymmetries, delayed vertical ground reaction force peaks, and mediolateral spatial instability ranked highest in importance. Reduced-feature models showed negligible performance loss, highlighting their parsimony and interpretability. Supervised machine learning algorithms can accurately distinguish true hemiplegic gait from mimicked patterns using a compact subset of gait features. The findings support data-driven, time-efficient gait assessments for clinical neurorehabilitation and for validating experimental protocols that rely on gait imitation. Full article

Non-equipotentiality in a grounding device can cause thermal heating in the screens of control cables that are grounded on both sides of high-voltage substations. At the same time, there is currently no approach for assessing the thermal endurance of cable screens that takes into account the configuration of the grounding device, the properties of the ground, and the connection. This paper presents a methodology for the experimental and computational determination of the thermal endurance of control cable shields in secondary circuits of 220–500 kV substations under short-circuit (SC) conditions. The method is based on full-scale imitation experiments using a sinusoidal current generator and verified numerical modeling in the ORU-M software. The potential and current density distribution in the cable shields were determined. The results showed that current densities in some circuits exceed permissible levels, confirming the risk of thermal damage. It was found that reconfiguring the grounding system—by densifying ground electrodes and increasing connections between grounding points—can reduce current density to acceptable values. The presented method allows for reliable assessment of the thermal endurance of cable shields without decommissioning the substation, making it suitable for the design and modernization of high-voltage facilities. Full article

I recently developed the concept of eco-cognitive openness and situatedness to explain how cognitive systems, whether human or artificial, engage dynamically with their surroundings to generate information and creative outcomes through abductive cognition. Human cognition demonstrates significant eco-cognitive openness, utilizing external resources like tools and cultural contexts to produce contextually rich hypotheses, sometimes highly creative via what I called “unlocked strategies.” Conversely, generative AI, such as large language models (LLMs) and image generators, employs “locked strategies,” relying on pre-existing datasets with minimal real-time environmental interaction—this leads to limited creativity. While these systems can yield some low-level degrees of creative outputs, their lack of human-like eco-cognitive openness restricts their ability to achieve high-level creative abductive feats, which remain a human strength, especially among the most talented. However, LLMs often outperform humans in routine cognitive tasks, exposing human intellectual limitations rather than AI deficiencies. Much human cognition is repetitive and imitative, resembling “stochastic parrots,” much like LLMs. Thus, LLMs are potent cognitive tools that can enhance human performance but also endanger creativity. Future AI developments, such as human–AI partnerships, could improve eco-cognitive openness, but risks like bias and overcomputationalization necessitate human oversight to ensure meaningful results. In collaborative settings, generative AI can serve as an epistemic mediator, narrowing the gap toward unlocked creativity. To safeguard human creativity, control over AI output must be maintained, embedding them in socio-cultural contexts. I also express concern that ethical and legal frameworks to mitigate AI’s negative impacts may fail to be enforced, risking “ethics washing” and “law washing.” Full article

With the advancement of UAV trajectory planning and sensing technologies, unmanned aerial vehicles (UAVs) are now capable of performing high-performance ground detection and search tasks. Mountainous regions, due to their complex terrain, have long been a focal point in the field of remote sensing. Effective UAV search tasks in such areas must consider not only horizontal coverage but also variations in detection range and angle caused by changes in elevation. Conventional algorithms typically require complete prior knowledge of the environment for trajectory optimization and often depend on scenario-specific policy models, limiting their generalizability. To address these challenges, this paper proposes a Feature-Aware Elite Imitation Multi-Agent Reinforcement Learning (FA-EIMARL) algorithm that leverages partial terrain information to construct a feature extraction network. This approach enables batch training across diverse terrains without the need for full environmental maps. In addition, an elite imitation mechanism has been proposed for convergence acceleration and task performance enhancement. Simulation results demonstrate that the proposed method achieves superior reward performance, convergence rate, and computational efficiency while maintaining strong adaptability to varying terrains. Full article

In traditional Chinese timber architecture, “column reduction” (Jian Zhu Zao) and “column relocation” (Yi Zhu Zao) enhances spatial continuity, yet often produces bending-dominated, material-intensive frames. This study develops a generative frame system that encodes raised beam logic into a parametric line-model workflow and couples it with simulation-based optimization. Informed by case analysis, the tool implements three lateral strategies—ridge-support revision, insertion of inclined members, and inclination of originally horizontal members—and one longitudinal strategy—longitudinal truss formation—whose use is governed by a user-defined historical authenticity parameter. Structural responses were evaluated using Karamba3D, and cross-section sizing was searched using Wallacei under gravity-dominant loading. The results indicate clearer load paths, greater axial-force participation, and reduced bending, yielding lower maximum displacements at comparable self-weight; moreover, the performance ranking aligns with the calibrated authenticity loss schedule, suggesting that the authenticity controller also acts as a practical proxy for expected stiffness gains. The framework improves design and modeling efficiency while offering quantitative decision support for culturally sensitive conservation and imitation design. Limitations include line-model idealization, simplified timber and joint behavior, gravity-only loading, and a modest historical corpus. The approach is extensible to other traditional systems via parameter and rule adaptation. Full article

Advances in AI and Brain–Computer Interfaces (BCIs) like Neuralink’s invite constructive Christian ethical responses that capitalize on these increasingly powerful technologies. This paper offers such a response. Its thought experiment partly draws on Immanuel Kant’s work Religion Within the Boundaries of Mere Reason in which he argues that the Son of God is the prototype of the perfectly good person and, as such, serves as the ideal model for anyone seeking to lead a moral life. Working within this Kantian framework, the anticipated capabilities of BCIs and AI could assist humans make moral progress and support their efforts to imitate the Son of God. These two technologies, coupled with a computer science approach to AI ethics known as Conditional Preference Networks, or CP-nets, offer a path forward. A case study in which a medical doctor with access to only one donor kidney must choose between two patients illustrates how BCIs and AI can help. Full article

In recent years, advancements in generative artificial intelligence have led to the development of sophisticated tools capable of mimicking diverse artistic styles, opening new possibilities for digital creativity and artistic expression. This paper presents a critical assessment of the style replication capabilities of contemporary generative models, evaluating their strengths and limitations across multiple dimensions. We examine how effectively these models reproduce traditional artistic styles while maintaining structural integrity and compositional balance in the generated images. The analysis is based on a new large dataset of AI-generated works imitating artistic styles of the past, holding potential for a wide range of applications: the “AI-Pastiche” dataset. This study is supported by extensive user surveys, collecting diverse opinions on the dataset and investigating both technical and aesthetic challenges, including the ability to generate outputs that are realistic and visually convincing, the versatility of models in handling a wide range of artistic styles, and the extent to which they adhere to the content and stylistic specifications outlined in prompts, preserving cohesion and integrity in generated images. This paper aims to provide a comprehensive overview of the current state of generative tools in style replication, offering insights into their technical and artistic limitations, potential advancements in model design and training methodologies, and emerging opportunities for enhancing digital artistry, human–AI collaboration, and the broader creative landscape. Full article

Unmanned aerial vehicles (UAVs) often face significant challenges in trajectory tracking within complex dynamic environments, where uncertainties, external disturbances, and nonlinear dynamics hinder accurate and stable control. To address this issue, a bio-inspired deep reinforcement learning (DRL) algorithm is proposed, integrating behavior cloning (BC) and long short-term memory (LSTM) networks. This method can achieve autonomous learning of high-precision control policy without establishing an accurate system dynamics model. Motivated by the memory and prediction functions of biological neural systems, an LSTM module is embedded into the policy network of the Twin Delayed Deep Deterministic Policy Gradient (TD3) algorithm. This structure captures temporal state patterns more effectively, enhancing adaptability to trajectory variations and resilience to delays or disturbances. Compared to memoryless networks, the LSTM-based design better replicates biological time-series processing, improving tracking stability and accuracy. In addition, behavior cloning is employed to pre-train the DRL policy using expert demonstrations, mimicking the way animals learn from observation. This biomimetic plausible initialization accelerates convergence by reducing inefficient early-stage exploration. By combining offline imitation with online learning, the TD3-LSTM-BC framework balances expert guidance and adaptive optimization, analogous to innate and experience-based learning in nature. Simulation experimental results confirm the superior robustness and tracking accuracy of the proposed method, demonstrating its potential as a control solution for autonomous UAVs. Full article

Deepfake audio refers to speech that has been synthetically generated or altered through advanced neural network techniques, often with a degree of realism sufficient to convincingly imitate genuine human voices. As these manipulations become increasingly indistinguishable from authentic recordings, they present significant threats to security, undermine media integrity, and challenge the reliability of digital authentication systems. In this study, a robust detection framework is proposed, which leverages the power of self-supervised learning (SSL) and attention-based modeling to identify deepfake audio samples. Specifically, audio features are extracted from input speech using two powerful pretrained SSL models: HuBERT-Large and WavLM-Large. These distinctive features are then integrated through an Attentional Multi-Feature Fusion (AMFF) mechanism. The fused features are subsequently classified using a NeXt-Time Delay Neural Network (NeXt-TDNN) model enhanced with Efficient Channel Attention (ECA), enabling improved temporal and channel-wise feature discrimination. Experimental results show that the proposed method achieves a 0.42% EER and 0.01 min-tDCF on ASVspoof 2019 LA, a 1.01% EER on ASVspoof 2019 PA, and a pooled 6.56% EER on the cross-channel ASVspoof 2021 LA evaluation, thus highlighting its effectiveness for real-world deepfake detection scenarios. Furthermore, on the ASVspoof 5 dataset, the method achieved a 7.23% EER, outperforming strong baselines and demonstrating strong generalization ability. Moreover, the macro-averaged F1-score of 96.01% and balanced accuracy of 99.06% were obtained on the ASVspoof 2019 LA dataset, while the proposed method achieved a macro-averaged F1-score of 98.70% and balanced accuracy of 98.90% on the ASVspoof 2019 PA dataset. On the highly challenging ASVspoof 5 dataset, which includes crowdsourced, non-studio-quality audio, and novel adversarial attacks, the proposed method achieves macro-averaged metrics exceeding 92%, with a precision of 92.07%, a recall of 92.63%, an F1-measure of 92.35%, and a balanced accuracy of 92.63%. Full article

Environmental protection tax (EPT), as a major tool to improve air quality and reduce carbon emissions, is of great significance for promoting urban low-carbon transformation. In this context, this paper has compiled a dataset from 282 Chinese cities during 2006–2022 and empirically identify the implication of EPT for carbon emissions at the city level by using the intensity difference-in-differences (I-DID) model. The result discloses that EPT greatly lowers carbon emissions by an average of 10.9% compared to non-pilot cities. Even after conducting some robustness checks, the result remains unchanged. Mechanism testing reveals that EPT curbs carbon emissions through enhancing energy utilization efficiency, fostering green technological advancements, and modernizing urban industries. Meanwhile, we show that EPT exerts a more substantial effect on carbon emissions in innovative cities, central and western cities, non-industrial-based cities, and non-resource-dependent cities. More importantly, EPT greatly promotes imitation and learning in neighboring regions, forming a radiation impact upon carbon reduction in surrounding areas. Hence, these results offer an important decision-making guide for optimizing the EPT system, strengthening the coordinated governance of carbon emission across regions, and ultimately promoting urban low-carbon development. Full article