Search Results (163)

This article evaluates the use of probing classifiers to modify the internal hidden state of a chess-playing transformer, which has been trained on sequences of chess moves and can generate new moves with prompted. Probing classifiers are a technique for understanding and modifying the operation of neural networks in which a smaller classifier is trained to use the model’s internal representation to learn a probing task. The aim of this research is to discover whether the learned model possesses an editable internal representation of the chess game, despite being trained without explicit information about the rules of chess. We contrast the performance of standard linear probes against Sparse Autoencoders (SAEs), a latent space interpretability technique designed to decompose polysemantic concepts into atomic features via an overcomplete basis. Our experiments demonstrate that linear probes trained directly on the residual stream significantly outperform probes based on SAE latents. When quantifying the success of interventions via the probability of legal moves, linear probe edits achieved an 88% success rate, whereas SAE-based edits yielded only 41%. These findings suggest that while SAEs are valuable for specific interpretability tasks, they do not enhance the controllability of hidden states compared to raw vectors. Finally, we show that the residual stream respects the Markovian property of chess, validating the feasibility of applying consistent edits across different time steps for the same board state. Full article

Laser Powder Bed Fusion (L-PBF) can enable in situ microstructural tailoring of metallic components by precisely controlling the layer-wise processing parameters. Layer rescanning is one such strategy used to induce localized microstructural modification. In this study, we investigated the effect of a lattice-based selective rescanning approach applied to different base scan strategies for Ti-6Al-4V samples. The lattice regions were selectively rescanned at 50% reduced laser power relative to the initial scan along the same laser path. Relative density, porosity, martensitic α′ morphology, phase fraction, and Vickers microhardness were compared with those of non-rescanned reference counterparts. Different scan strategies, including unidirectional, stripes, and chess, exhibited distinct responses to selective rescanning, resulting in localized variations in martensitic phase formation and hardness values. The extent of localized microstructural modification and hardness enhancement was strongly governed by the underlying scan strategy. Selective rescanning using the stripes strategy yielded the largest contrast between non-rescanned and rescanned regions. The unidirectional strategy showed strong effects of rescanning, but the heat-affected zones extended to the non-rescanned regions. In contrast, the chess strategy exhibited comparatively moderate changes owing to its inherent thermal-management characteristics. These findings demonstrate that selective rescanning can provide an effective, localized approach for tailoring microstructure and hardness enhancement in L-PBF Ti-6Al-4V, with its effectiveness strongly dependent on the underlying scan strategy. Full article

Teacher motivation plays a critical role in the successful implementation of curriculum enrichment programmes, yet it remains underexplored in many educational initiatives. The study contributes insights into teacher motivation in early childhood education and offers practical guidance for strengthening the sustainability of enrichment programmes through integrated recognition, support, and incentive structures. This study investigates the impact of rewards and recognition on teachers’ engagement in the Tsogo Sun Moves for Life chess programme in early childhood education classrooms within the King Cetshwayo District, South Africa. A qualitative case study design was used, with data collected through semi-structured interviews, observations, and document analysis, and analysed using thematic analysis. The findings indicate that while teachers valued teaching resources, coordinator support, and certificates of appreciation, these forms of recognition were insufficient to sustain long-term engagement. Teachers emphasised the need for meaningful acknowledgment and tangible incentives to justify the additional workload associated with programme implementation. Drawing on Self-Determination Theory and Herzberg’s Two-Factor Theory of Motivation, the study highlights the interaction between intrinsic and extrinsic motivation in shaping teachers’ commitment. Full article

Human adenoviruses (HAdVs) remain prominent global human pathogens, particularly in dense, crowded populations. The advent of genomic and bioinformatic tools allows for high-resolution means to identify, characterize, and understand these pathogens. These tools also provide the basis for the standardization of names, as well as an accessible archive of all genotypes (“Human Adenovirus Working Group”). This overview and perspective of all the genotypes in one setting provides a better understanding of the mechanisms of their molecular evolution: genome recombination plays a major role in the emergence of novel adenoviral pathogens. In the context of the fidelity of their DNA polymerase replication machinery, this strategy provides entry into immune-naïve host populations through the acquisition of genome sequences that may include antigenic epitopes that have not circulated commonly, widely, or recently, as well as sequences encoding host cell entry proteins. Using the “chess” metaphor for describing the rapid evolution of RNA viruses, we propose a similar but diametrically opposed “White King Reigns in the Family of Human Adenoviruses”. Full article

The rapid growth of online chess has intensified the challenge of distinguishing engine-assisted from authentic human play, exposing the limitations of existing approaches that rely solely on deterministic evaluation metrics. This study introduces a proof-of-concept hybrid framework for discriminating between engine-like and human-like chess play patterns, integrating Stockfish’s deterministic evaluations with stylometric behavioral features derived from the Maia engine. Key metrics include Centipawn Loss (CPL), Mismatch Move Match Probability (MMMP), and a novel Curvature-Based Stability (ΔS) indicator. These features were incorporated into a convolutional neural network (CNN) classifier and evaluated on a controlled benchmark dataset of 1000 games, where ‘suspicious’ gameplay was algorithmically generated to simulate engine-optimal patterns, while ‘clean’ play was modeled using Maia’s human-like predictions. Results demonstrate the framework’s ability to discriminate between these behavioral archetypes, with the hybrid model achieving a macro F1-score of 0.93, significantly outperforming the Stockfish-only baseline (F1 = 0.87), as validated by McNemar’s test (p = 0.0153). Feature ablation confirmed that Maia-derived features reduced false negatives and improved recall, while ΔS enhanced robustness. This work establishes a methodological foundation for behavioral pattern discrimination in chess, demonstrating the value of combining deterministic and human-centric modeling. Beyond chess, the approach offers a template for behavioral anomaly analysis in cybersecurity, education, and other decision-based domains, with real-world validation on adjudicated misconduct cases identified as the essential next step. Full article

Online chess platforms generate vast amounts of game data, presenting opportunities to analyze match outcomes using machine learning approaches. This study develops and compares four machine learning models to classify chess game results (White win, Black win, or Draw) by integrating player rating information with game dynamic metadata. We analyzed 11,510 complete games from the Lichess platform after preprocessing a dataset of 20,058 initial records. Seven key features were engineered to capture both pre-game skill parameters (player ratings, rating difference) and game complexity metrics (game duration, turn count). Four machine learning algorithms were implemented and optimized through grid search cross-validation: Multinomial Logistic Regression, Random Forest, K-Nearest Neighbors, and Histogram Gradient Boosting. The Gradient Boosting classifier achieved the highest performance with 83.19% accuracy on hold-out data and consistent 5-fold cross-validation scores (83.08% ± 0.009%). Feature importance analysis revealed that game complexity (number of turns) was the strongest correlate of the outcome across all models, followed by the rating difference between opponents. Draws represented only 5.11% of outcomes, creating class imbalance challenges that affected classification performance for this outcome category. The results demonstrate that ensemble methods, particularly gradient boosting, can effectively capture non-linear interactions between player skill and game length to classify chess outcomes. These findings have practical applications for chess platforms in automated content curation, post-game quality assessment, and engagement enhancement strategies. The study establishes a foundation for robust outcome analysis systems in online chess environments. Full article

Background: The clinical utility of risk scores predicting shunt dependency after aneurysmal subarachnoid hemorrhage (aSAH) remains limited due to scarce validation data. This multicenter pooled analysis aimed to assess the predictive accuracy of existing post-aSAH shunt risk scores. Methods: Consecutive aSAH cases treated at two German university hospitals from January 2010 to July 2023 were pooled into a validation cohort. Total scores for the CHESS, CHESS-Huckman, and SDASH risk models were calculated, and their diagnostic performance was compared using receiver operating characteristic (ROC) curve analysis. Results: A total of 813 patients were included, of whom 215 (26.4%) required ventriculoperitoneal shunt placement within a median time of 29 days post-aSAH. All three risk scores were significantly associated with shunt dependency. ROC analysis showed that the CHESS-Huckman score had the highest predictive accuracy (AUC: 0.792, 95% CI: 0.761–0.824), followed by the SDASH (AUC: 0.782, 95% CI: 0.750–0.814) and CHESS (AUC: 0.780, 95% CI: 0.748–0.812) scores. Pairwise comparisons of AUCs were not statistically significant. All three scores showed good overall calibration, with CHESS–Huckman performing best, as confirmed by calibration intercepts and slopes, Brier scores, and decile-based analysis. Higher CHESS–Huckman scores correlated with earlier shunt placement, whereas delayed shunting (>30 days after aSAH) was most common in patients with moderate CHESS–Huckman scores (7–8 points), occurring in 47.4% of cases compared to 41.4% and 33.3% in patients scoring 0–6 and 9–10 points, respectively. Conclusions: This multicenter analysis validated existing risk scores for predicting shunt dependency after aSAH, with the CHESS–Huckman score demonstrating the nominally highest diagnostic accuracy. Integrating these risk scores into clinical practice could enhance early identification of patients requiring shunting, potentially reducing external ventricular drain weaning time, shortening hospital stays, and lowering the risk of cerebrospinal fluid infections. Full article

The application of artificial intelligence through the brain–computer interface (BCI) is proving to be one of the great advances in neuroscience today. The development of surface electrodes over the cortex and very fine electrodes that can be stereotactically implanted in the brain have moved the science forward to the extent that paralyzed people can play chess and blind people can read letters. However, the introduction of foreign bodies into deeper parts of the central nervous system results in foreign body reaction, scarring, apoptosis, and decreased signaling. Implanted electrodes activate microglia, causing the release of inflammatory factors, the recruitment of systemic inflammatory cells to the site of injury, and ultimately glial scarring and the encapsulation of the electrode. Recordings historically fail between 6 months and 1 year; the longest BCI in use has been 7 years. This article proposes a biomolecular strategy provided by angiogenic cell precursors (ACPs) and nerve cell precursors (NCPs), administered intrathecally. This combination of cells is anticipated to sustain and promote learning across the BCI. Together, through the downstream activation of neurotrophic factors, they may exert a salutary immunomodulatory suppression of inflammation, anti-apoptosis, homeostasis, angiogenesis, differentiation, synaptogenesis, neuritogenesis, and learning-associated plasticity. Full article

Cancer patients use cannabis for medicinal purposes; however, few studies have examined hematology patients’ use of cannabis and no research to our knowledge has investigated the use of cannabis amongst hematology patients before and after hematopoietic stem cell transplant (HCT). The purpose of this longitudinal survey study was to assess aspects of cannabis use in patients who underwent HCT. Eligible patients (N = 30) completed two surveys before and 90 days following their HCT. The surveys inquired about several aspects of cannabis (e.g., rate of use, beliefs and attitudes, access to information) and physical and psychological outcomes (e.g., anxiety, comorbidities, graft-versus-host-disease). Rates of cannabis use decreased following HCT (n = 14, 46% to n = 11, 40%). Conversations on cannabis that were initiated by an oncology health care provider increased post-transplant (n = 3, 10% to n = 11, 37%). This coincided with fewer who were smoking cannabis as a primary consumption method (n = 5, 38 to n = 2, 18) and an increase in the use of pharmaceutical cannabinoid products (n = 4, 13% to n = 6, 21%) as well as oils and topicals. Of the total sample, 63% (n = 17) experienced post-treatment complications and 33% (n = 10) developed GVHD, six of whom where recent cannabis users. This study provided insight into cannabis use amongst HCT patients and warrants further research with this population, including more exploration of the relationship between GVHD and cannabis. Full article

Background/Objectives: People living with a spinal cord injury (PwSCI) present numerous complications at a systemic level that negatively impact their physical and mental health as well as their quality of life. The purpose of this study was to describe the clinical profile of PwSCI living in nursing homes (NHs), Complex Continuing Care Systems (CCCs), home care (HC), and inpatient mental health facilities (MHs) in nine Canadian provinces and territories. Methods: We analyzed data collected with the following assessment tools: Resident Assessment Instrument (RAI) Minimum Data Set (RAI-MDS 2.0), RAI–MH, RAI-HC, Cognitive Performance Scale, Activities of Daily Living (ADL) Hierarchy Scale and impairments in instrumental ADLs (IADLs), Pain Scale, Changes in Health, End-Stage Disease, Signs, and Symptoms (CHESS) Scale, Depression Rating Scale, and Deafblind Severity Index (DBSI). We reported counts (n) and percentages (%) and performed Chi-square tests with a Bonferroni correction to determine the statistical significance of the differences in frequencies within and between care settings. Results: We identified 13,136 PwSCI, predominantly males and younger than comparison groups. PwSCI presented fewer comorbidities but reported higher pain than comparison groups. Almost all of the PwSCI in NHs (99.4%) and CCCs (98.9%) needed assistance to perform ADLs. Conclusions: The prevalence of comorbidities and impairments following SCI varies based on the clinical setting. The present clinical profile of PwSCI will inform interventions to improve health of PwSCI across the continuum of care. Full article

Online gamers have increased exponentially in the last few years in all types of online games, including mind-sport games. These games, like Bridge or Chess, have been traditionally played face-to-face. Nowadays more and more players prefer to use online platforms to play mind-sport games. Previous studies have investigated different aspects of online games and even a few on mind-sport games. However, the frameworks WOM (Word-of-Mouth) and SCT (Social Cognitive Theory) have been sparsely used in this context. In this manner, the present article proposes two objectives: (1) using the SCT in order to analyse the impact of the sociological factor on user engagement in mind-sport online games and (2) analysing how the WOM affects user engagement in mind-sport online games. Specifically, the proposed PLS-SEM model is defined by combining five constructs from these frameworks: (1) health consciousness, (2) WOM and emotional behaviour, (3) self-efficacy, (4) cognitive engagement, and (5) behavioural intention. The findings reveal that health consciousness affects WOM and emotional behaviour in a positive way as players desire well-being. Also, WOM and emotional behaviour affect cognitive engagement, as positive comments encourage high-skill gamers in mind sports. Finally, this study shows how the environmental factor of SCT is represented by WOM and emotional behaviour in an indirect way and the personal factor represented by self-efficacy in a direct way to positively influence behaviour intention. Full article

This study investigates the cognitive processes underlying chess expertise by examining planning, cognitive reflection, inter-temporal choice, and risky choice in chess players. The study involves 25 chess players and 25 non-chess players, comparing their performance on the Tower of London (TOL) task, Cognitive Reflection Test (CRT), inter-temporal choice (ITC), and risky choice tasks. Results indicate that chess players outperform non-chess players in TOL and CRT, showing superior planning and cognitive reflection abilities. Chess players also prefer future rewards over immediate ones in ITC, suggesting a higher propensity for future more rewarding options. In risky choice tasks, chess players made more decisions based on expected value than non-chess players, but the evidence in favour of differences between groups is very weak. Despite this study not being able to establish causality, the findings highlight the cognitive advantages associated with chess expertise and suggest potential areas for further research on the transfer of cognitive skills from chess to other domains and differences in general abilities between experts and novices. Full article

This study introduces a novel approach to chess analysis based on center-of-mass dynamics and discrete fractal derivatives, offering an alternative framework for evaluating gameplay strategies. Unlike conventional methods that rely on exhaustive search and statistical simulations, our model provides a macroscopic perspective by analyzing the collective motion of pieces over time. By representing chess positions as a dynamic system in ${\mathbb{R}}^2$, we identify key movement patterns—such as oblique, parallel, and orthogonal trends—revealing strategic tendencies throughout the game. Additionally, fractal derivatives enable the detection of subtle momentum shifts and long-term imbalances, enhancing the understanding of decision-making processes. This approach is computationally efficient and adaptable, extending beyond chess to applications in sports analytics and real-time strategy games. These findings highlight the potential of interdisciplinary techniques in capturing complex strategic behavior within dynamic environments. Full article

Introduction: Deep reinforcement learning (deep RL) integrates the principles of reinforcement learning with deep neural networks, enabling agents to excel in diverse tasks ranging from playing board games such as Go and Chess to controlling robotic systems and autonomous vehicles. By leveraging foundational concepts of value functions, policy optimization, and temporal difference methods, deep RL has rapidly evolved and found applications in areas such as gaming, robotics, finance, and healthcare. Objective: This paper seeks to provide a comprehensive yet accessible overview of the evolution of deep RL and its leading algorithms. It aims to serve both as an introduction for newcomers to the field and as a practical guide for those seeking to select the most appropriate methods for specific problem domains. Methods: We begin by outlining fundamental reinforcement learning principles, followed by an exploration of early tabular Q-learning methods. We then trace the historical development of deep RL, highlighting key milestones such as the advent of deep Q-networks (DQN). The survey extends to policy gradient methods, actor–critic architectures, and state-of-the-art algorithms such as proximal policy optimization, soft actor–critic, and emerging model-based approaches. Throughout, we discuss the current challenges facing deep RL, including issues of sample efficiency, interpretability, and safety, as well as open research questions involving large-scale training, hierarchical architectures, and multi-task learning. Results: Our analysis demonstrates how critical breakthroughs have driven deep RL into increasingly complex application domains. We highlight existing limitations and ongoing bottlenecks, such as high data requirements and the need for more transparent, ethically aligned systems. Finally, we survey potential future directions, highlighting the importance of reliability and ethical considerations for real-world deployments. Full article

Steganography is a technique for hiding secret information in imperceptible carriers and transmitting it. Unlike traditional embedding-based steganography, generative steganography can generate stego-carriers directly from secret messages, thus avoiding modifications to natural carriers that steganalysis can detect. As a branch of generative steganography, game-behavior-based steganography transmits secret information by encoding game behavior. It can naturally integrate with real interaction scenarios, exhibiting strong concealment and undetectability. To this end, this paper proposes a generative steganography based on Chinese Chess record construction. Firstly, an AlphaZero model was trained to achieve a high level in Chinese Chess, then transmit secret information by encoding chess behavior. Specifically, in each chess step, the model generates all the current feasible moves and encodes the moves that meet the threshold strategy according to probability. Then, the appropriate move will be selected according to the secret information. To ensure the reasonableness of the generated chess records, this paper controlled the game process and designed a database of fixed opening chess records. The proposed method can hide an average of 413 bits of information for each carrier and effectively resist common image attacks. Regarding anti-steganalysis, the proposed method achieved accuracy rates of 0.498 and 0.497 on XuNet and YeNet, respectively, outperforming other behavior-based steganography techniques. Full article