We present a serious game framework that challenges players to optimize flood rescue operations by manipulating six parameters governing boat evacuations and survival kit distribution. The underlying model uses competitive Lotka-Volterra models where rescue mechanisms act as “predators” reducing flood deaths (the “prey”). Players navigate a six-dimensional parameter space to minimize both deaths and costs; they make real-time decisions under uncertainty. Our game implements pseudo-random fluctuations using differences between nontrivial Riemann Zeta zeros — sequences that appear random to human perception while containing non-obvious structure. This design choice aims to create realistic uncertainty that engages rather than frustrates players. The game addresses a critical gap in municipal emergency planning: identifying effective rescue strategies before flood disasters strike. By crowdsourcing optimization through competitive gameplay, we transform abstract mathematical modeling into engaging challenges while generating actionable policy recommendations. Players receive immediate visual feedback showing how their parameter choices affect survival curves and resource allocation over a simulated 10-day flood event. This work contributes to serious games research by demonstrating how complex multi-objective optimization in crisis scenarios can be gamified and explores how pseudo-random sequences may create more intuitive challenge experiences than true randomness.

The first phase of qualitative research is often hindered by the time-consuming process of transcribing and analyzing low-quality audio interviews. This paper introduces a novel human-in-the-loop pipeline that leverages Multimodal Large Language Models (MLLMs) and Retrieval-Augmented Generation (RAG) to overcome these. Our case study focuses on flood-related interviews conducted in North Bohemia, Czech Republic, in Czech, a language for which many AI models have not yet been optimized. Our proposed methodology employs MLLMs for both generating quantifiable, emotionally nuanced interview stimuli and for reconstructing interview transcripts from poor audio recordings. The initial AI-generated transcriptions are then refined and validated by a human, ensuring high accuracy and contextual integrity. Subsequently, a RAG system is utilized for efficient and transparent thematic analysis of the corrected transcripts. Our results demonstrate that this hybrid approach can achieve over 99

Medical personnel are one group for which simulations and VR training make the highest impact on the actual well-being of the population. Most important is that the simulation has real bases and all information gained in scenarios are usable in real life. TBI (traumatic brain injury) is a very serious medical condition requiring intense medical care. Survival likelihoods, which we estimate, are chillingly low. We investigate whether there is an interplay between various categorical variables (biological sex, mechanisms and locations of accident, trauma mechanism, alcohol intoxication, illicit drug consumption, as well as ocular, verbal, motoric and pupil responses, and survival or death). In the medical scenario of the emergency room into which TBI victims are delivered, physicians habitually convert degrees of injury or neurological impairment severity into cardinal numbers that are then used to compute an index. Because this method is statistically fallacious, we include Bayesian methods that prevent this fallacy. The sequence of analysis steps involves one-hot encoding, dimension reduction, kernel density estimation, clustering algorithms, followed by computing the off-diagonal entries in the confusion matrix, which are a significance measure. We generate heat maps of different clusters to identify how survival rates relate to the categorical variables that characterize the severity of—and mechanisms that caused—TBI. In the realm of serious gaming, the development of realistic and accurate TBI simulation logic necessitates the integration of actual patient data. We discuss the general implications, as well as warnings and cautions when creating simulations as serious games.

Anticipating extreme events in time-series data — whether financial crashes, river floods, or solar activity disruptions — is essential for effective decision-making yet complicated by the difficulty of distinguishing genuine trends from random fluctuations. We present a parameter-free diagnostic framework based on harmonic numbers and record-breaking event statistics that assesses whether detrended time series retain exploitable structure or are consistent with randomness. By computing first differences and analyzing the frequency of record-high and record-low events separately, the method detects directional asymmetries not detectable by symmetric volatility models. We validate the framework across four empirical domains: 72 NYSE equity securities (2001–2024), historical water levels of the Nile River (861–1111 ad), recent water levels of the River Elbe (1995–2024), and sunspot counts (1610–1995 ad). Results show that approximately 80

This paper presents Citizen2Vec, a serious gaming approach designed to provide decision-makers with an understanding of the complex interplay of environmental, social, and health factors within an urban population. As part of the ongoing “Blood Drop Project” study conducted in Ústí nad Labem, Czech Republic, Citizen2Vec employs a novel approach, analogous to doc2vec, by vectorizing individual citizen’s properties (data collected from surveys, physiological measurements, and geospatial mapping). Vectorization generates a comprehensive digital representation of the population, enabling to-be-discovered relationships between health, living conditions, and environmental factors. Citizen2Vec serves as both a predictive simulation tool and a decision-support system, allowing policymakers to model the effects of urban policies on public health and social well-being. By tuning parameters related to air quality, infrastructure, healthcare, and social support systems, the game outputs the impact on the virtual/simulated citizenry. Furthermore, it provides insights into how different interventions vary in effectiveness. The competition via gaming provides proposals for actual, real modifications for the disadvantaged in society by informing how effective putative interventions can be and which are most attractive because subsequent costs are (literally) part of the game. The game’s development integrates advanced analytical techniques, namely dimension reduction and computed confusion matrices. These enable the identification of specific patterns (likelihood functions) and associations between environmental exposures, living and socio-economic conditions, and health issues. Citizen2Vec represents a significant advancement in Human-Computer Interaction (HCI) for public health planning. The advantage of crowd gaming is to invite gamers to provide decision-makers with discovered policy scenarios and these choices can promote the well-being of the most disadvantaged members of an urban population.

This study critically evaluates the reliability and validity of the widely used Five Factor Model (FFM) or 'Big Five' personality traits framework across 12 Latin American countries. Conventional psychometric assessments based on factor analysis have significant methodological limitations when applied to categorical data. Addressing these concerns, we employed a Bayesian statistical approach utilizing Dirichlet and Beta distributions for categorical responses obtained from 5,175 participants who completed the IPIP-R questionnaire. Our novel methodology includes Monte Carlo simulations, confusion matrices, and probability density function estimations, effectively compensating for inherent sample size imbalances. Findings demonstrate substantial cultural variations in the distribution of personality traits, contradicting the presumed universality of the FFM. Additionally, notable differences were observed between male and female respondents, influenced by nationality. Furthermore, natural language processing techniques combined with the UMAP dimensionality reduction algorithm revealed that linguistic clustering of questionnaire items does not explain cultural differences. Our results demonstrate the inadequacy of factor analysis for analyzing categorical psychometric data, necessitating instead rigorous Bayesian methods. This study significantly impacts how personality assessments should be utilized in policymaking, corporate environments, and artificial intelligence applications, emphasizing the necessity of culturally sensitive and statistically robust approaches. The outcomes are discussed in relation to creation of citizen profiles and advanced personality modelling.

In recent years, the field of artificial intelligence (AI) has witnessed significant advancements, particularly with the development of transformer architectures and pre-learned feature extraction techniques. Before this, thousands of high-quality scientific papers using biological landmarks (manually placed by a researcher) on faces have identified relations between facial morphology and various human ratings of them. Geometric Morphometrics provides a framework for quantifying and analyzing shape variations in biological organisms, allowing for the extraction of morphological features that rely on landmark identification. In avatar creation, matters are more complex. Game designers, when attempting to mimic anticipated ratings, oftentimes cannot rely strictly on landmarks. As we document here, we can relax the rigor of constructed, geometrically identifiable landmarks and replace them with characteristic points identified by artificial neural networks (aNNs), obviating the need for tiresome, meticulous landmark identification by human operators. By integrating these features into the avatar generation process, we can ensure that the resulting avatars exhibit desired characteristics, while still adhering to biological constraints. We relate these characteristic points on female face images whose physical attractiveness has been rated by 50 males and 50 females (on a 7-point scale). We construct Dirichlet distributions of the ratings of each face and use these to investigate whether there are trends or biases in face ratings by male raters versus female raters. We use the coordinates of characteristic points on the faces to identify clusters of attractiveness and relate these clusters to heat maps of the ratings. We explore localized pseudo-symmetries, and howthey relate to these ratings. Our proposed system aims to leverage the strengths of transformer architectures and pre-trained neural networks, notably their ability to capture contextual information, to enhance the realism and biological accuracy of the avatars.

UFC (Ultimate Fighting Championship) video games aim to provide a realistic virtual experience of MMA (Mixed Martial Arts) fights. However, despite the seeming simplicity of the point-based scoring system used during these fights, judges’ holistic verdicts of identifying the winner often fail to accurately reflect the performances because, we find, they do not rely on the point-awarding system used during the fights. We investigate the extent to which judges employ holistic approaches in their scoring decisions, even though they are consistent in applying them. To overcome the shortcomings of the holistic approach, we propose a merit-based set of rules (based on the Bayesian statistical approach that is driven by the points accrued in each round of the fight) for identifying the winners of matches and tournaments. Merit-based sets of rules are commonplace in economic games, in which strategic decision-making and performance evaluation are common, and such rules infer that quantified measures do objectify the assessment of player/fighter performance in a far more satisfactory (and non-controversial) manner. We thereby emphasize the importance of significant strikes being a key component of scoring. Our analysis of data from 3039 UFC decision fights demonstrates that the potential of objective scoring methods will make tournaments more attractive for not only gamers but also spectators. Likewise, such objective evaluation methods are useful in economic games.

The debate about who is the GOAT (the “Greatest Of All Time”) is not restricted to Formula One competition. There are several possible criteria, other than the number of World Championships won. We approach this issue of attempting to identify the GOAT (assuming that there is only one), by looking at the race performance of 12 drivers who participated in Grand Prix races since 1950. (a) Unquestionably, the more race participation, the greater the chance of accumulating wins and therefore improving a driver’s ranking towards being the GOAT. (b) A similar line of argument occurs for the ranking via podiums. (c) The system of awarding points for ranks after the race has changed five times since 1950. Comparing points accrued by different drivers therefore necessitates reassigning points to achieve comparability. We also consider how rankings change if other point systems are used. (d) Bayesian statistical analysis allows not only the evaluation of who is eligible for GOAT without using a point system (via Beta and Dirichlet distributions) but can also estimate the probability that two contenders’ performances differ only by chance. Formula One also takes place in video games, particularly in tournaments. Here we provide evidence that avoiding point systems to identify winners of the tournament is advisable. We provide suspense by not revealing who is the GOAT in this Abstract. We provide the names of one or several GOATs (there might be several) in this publication.

The realm of retail investment encompasses a spectrum of decisions made by individuals seeking to allocate resources in anticipation of future returns. This study examines the implications of portfolio management and stock request behavior for developer- and economy-oriented game design. The research methodology uses a simulated investment game to bridge the gap between simulated environments and real-life investment scenarios. A total of 888 participants, aged 20–24 years, engaged in the game, which featured 66 companies and 49 rounds, simulating a 49-day tenure of companies in the market. We employed Bayesian statistical methodology to estimate the likelihood of participants being female based on the sex ratio in the sample. Additionally, we used SVD (singular value decomposition) to smooth the data matrix representing the investment choices made by participants across the companies and rounds. Furthermore, we employed the DBSCAN algorithm to detect clustering in the 66 rows of SVD-smoothed mean share prices. By incorporating elements of realism, complexity, and strong participant motivation, we provided a robust and immersive investment experience compared to existing investment games. The findings contribute to a better understanding of individual investment decision-making processes within the game theory framework.

The Wilhelm Scream is the sound effect that first appeared in 1951 and has been used in an almost uncountable number of movies and video games as an “Easter egg” related to pain, injury, and displeasure. The Intensity Paradox theory claims that vocalizations are rated valence-wise with very low accuracy. The learning hypothesis, on the other hand, suggests that repeated exposure allows for overcoming the intensity paradox. To test this, we collected a large sample of two ratings each by 902 raters and developed a novel statistical approach, based on Bayesian statistics, in which we determine the maximum-likelihood probability of Beta distributions of these responses. We asked our participants to rate the Wilhelm Scream mixed in with the other high-intensity vocalizations. The outcome showed an unexpected result—an extremely high precision of rating and consistency in repeated exposure. Furthermore, older men (but not women) rated with lower consistency. This provides, using a natural experiment, novel support for the learning effect of vocalization and the importance of using Easter eggs as a vehicle of familiarity related to gaming attractiveness and audiovisual media creation.

In the gaming world, as well as in assistive technologies, profile creation and further characterizations of humans are basic. Oftentimes, Likert-type scales are needed to collect responses in questionnaires. A subsequent mistake by the analyst occurs when he/she computes the composition score derived from these response categories. Because: such a score can only be computed if these categorical responses are converted into computable numbers. An AI method, namely the application of artificial neural networks, can extract information and overcome various erroneous statistical methods (such as: linearizing responses of Likert scales, computing scores and correlations, and disregarding the ‘curse of dimensionality’). We collected data from 480 respondents who were asked to specify boundaries between the colors of the blackbody spectrum (‘rainbow’). We first used an auto-encoder for dimension reduction, then searched for categories by implementing clustering algorithms, computed likelihood plots, and calculated confusion matrices based on Dirichlet distributions. We found that every respondent was a member of one of only three clusters. Each cluster is characterized by a different distribution of the color boundaries between purple-blue, blue-green, and so on. Surprisingly, the boundaries between some colors in one cluster where within the color interval for members of another cluster. Conclusion: where people see a color boundary is far from obvious even within a shared range. There are implications for several fields that have been using ‘Stone Age statistics’ involving scores, etc. as listed above, among them psychology, sociology, behavioral economy, and human-computer interaction. The modern approach, based on AI, should be adopted by researchers in these fields to ensure reproducibility and provide insight into participant/user/citizen profiles.

Objectives. Research of sexuality often times includes individuals who choose to be researched and are willing to undergo the procedures. Samples of population (students and self-selected enthusiasts) may not resemble the nation's population structure nor general idea of random hu-man even though the psychological mechanisms researched are thought to be universal. The aim of this study was to investigate the relationship between proximal measures of domain specific and non-specific excitation and inhibition in two (probabilistic and non-probabilistic) samples. Sample and settings. The authors collected ap-proximately 400 individuals from Central European country - Czech Republic - via specialized agency to match the probabilistic sample (stratified sample) and a second similarly sized sample (chance sample) via snowball sampling and online advertisements at the same time. The participants were asked to fill in two question-naires widely used by sex researchers (Sexual Inhibition and Excitation Scale, and Sociosexual Orientation Inventory Revised), and one scale designed to map broader motivations to avoid aversive outcomes and motivation to approach goal-oriented outcomes (Behavioral Inhibition and Activation Scale). Hypotheses. The authors expected partial or complete confirmation of previously found results (general and sexuality specific activation and inhibition and sociosexuality), and also that the replication will be affected by the sample choice. Statistical analyses. To follow methodologies of previous studies used to create hypotheses the authors decided to use partial correlation con-trolling for age of the participants.Results. Generally, the chance sample findings resembled the published results whereas the stratified sample showed bigger differences. The results, relying on correlational analyses as majority of studies does, provided a valuable in-sight to impact of sampling on results. Using the SEM methodology the study provided further support for the incomparability of the results obtained employing different sampling.Limitations. The sample size, thus larger than usual studies, could be higher.

In real-world scenarios, humans estimate a large proportion of their perceived world contextually and use previous information to adjust or modify their expectations and responses. A typical example of congruence is when a smile (visual modality) is accompanied by a laugh (acoustic modality). Pain and pleasure are extremely intensive affects, rarely correctly assessed. It rarely happens that affective communication is incongruent in different modalities. Intentional combinations of these two affective expressions may be implemented by using audiovisual media. It is to be expected that presentations involving incongruent combinations during gaming and other interactions with machines will alter perceptions and the impact of ambiguity. To evaluate the impact of sensory crosstalk a novel statistical analysis was developed to estimate the impact of each modality. The results show that the visual modality dominates for the pairing pain/pleasure for. The acoustic modality dominates for pain/pleasure, pain/neutral, and pleasure/neutral. For neutral/pain and neutral/pleasure, neither the visual nor the acoustic modality enables highly correct ratings. The findings are of high value to the psychology of perception on a theoretical level and game/media developers as application fields.

In our everyday life we rely on set of heuristics that involve estimation of meaningful connections between events. In human evolutionary history, it was less costly to overestimate the meaning. The psychological phenomenon of apophenia (overperception of patterns) is then an adaptive response. It may manifest also as overperception of visual patterns (pareidolia). The underperception was rarely studied and researchers mainly used unsuitable stimuli sets for the purpose. After researching this phenomenon using patterns with transparency, geometric shapes, and color (Boschetti et al., 2023 ), we developed new set of black and white high-contrast stimuli. These were presented to participants four times in different orientations to limit guessing to 6% chance. Using ANN (Artificial Neural Networks), we associated the responses to the Rational Experiential Multimodal Inventory Subscales Rationality and Intuition. We were able to identify two clusters for each subscale and found associations of the participant responses with pattern identification success (or lack thereof). Our discoveries extend previous findings concerning this phenomenon and provides us with a foundation for constructing and designing artificial environments with high attention to cues given to users.

(1) Background: In addition to verbalizations, facial expressions advertise one’s affective state. There is an ongoing debate concerning the communicative value of the facial expressions of pain and of pleasure, and to what extent humans can distinguish between these. We introduce a novel method of analysis by replacing human ratings with outputs from image analysis software. (2) Methods: We use image analysis software to extract feature vectors of the facial expressions neutral, pain, and pleasure displayed by 20 actresses. We dimension-reduced these feature vectors, used singular value decomposition to eliminate noise, and then used hierarchical agglomerative clustering to detect patterns. (3) Results: The vector norms for pain–pleasure were rarely less than the distances pain–neutral and pleasure–neutral. The pain–pleasure distances were Weibull-distributed and noise contributed 10% to the signal. The noise-free distances clustered in four clusters and two isolates. (4) Conclusions: AI methods of image recognition are superior to human abilities in distinguishing between facial expressions of pain and pleasure. Statistical methods and hierarchical clustering offer possible explanations as to why humans fail. The reliability of commercial software, which attempts to identify facial expressions of affective states, can be improved by using the results of our analyses.

Objectives: Research of sexuality often times includes individuals who choose to be researched and are willing to undergo the procedures. Samples of population (students and self-selected enthusiasts) may not resemble the nation’s population structure nor general idea of random human even though the psychological mechanisms researched are thought to be universal. The aim of this study was to investigate the relationship between proximal measures of domain specific and non-specific excitation and inhibition in two (probabilistic and non-probabilistic) samples. Sample and settings: We collected approximately 400 individuals from Central European country – Czech Republic – via specialized agency to match the probabilistic sample and a second similarly sized sample via snowball sampling and online advertisements at the same time. The participants were asked to fill in two questionnaires widely used by sex researchers (Sexual Inhibition and Excitation Scale and Sociosexual Orientation Inventory), and one scale designed to map broader motivations to avoid aversive outcomes and motivation to approach goal-oriented outcomes (Behavioral Inhibition and Activation Scale). Hypotheses: We expected partial or complete confirmation of previously found results (general and sexuality specific activation and inhibition and sociosexuality). We also expected that the replication will be affected by the sample choice. Statistical analyses: To follow methodologies of previous studies used to create hypotheses we decided to use partial correlation controlling for age of the participant. Results: Generally, the non-probabilistic sample findings resembled the published results whereas the probabilistic sample differed, especially in men. Furthermore, findings from women were the more consistent of the two sexes when comparing the probabilistic and non-probabilistic sample outcomes. Our results, relying on correlational analyses as majority of studies does, provide a valuable insight to impact of sampling on results. Limitations: The sample size, thus larger than usual studies, could be higher. We have been unable to research latent variables due to methodological constrains of the method - correlation.

As software technology becomes more and more present in our lives, its use has arched from purely purposeful (such as complex computations), to purely fun-related (such as video-games). We can combine both in so-called ‘serious games’. These can serve as means of training, instructing, tutoring, teaching, and also as a basis for data collection in various scenarios. Data collection and analysis is the primary one we pursue in our proof-of-concept presentation. Our ‘serious gaming’ is the gamification of an environment-related challenge. We intend to use the Serious Game for Energy Science scenario and populate it with formalized task models. Gamers will be exposed to videos that could modify their behavior. Graphically, the game will not be very complex so as to keep the focus on input information. The environment will be visualized in five levels, from bucolic to devastated. Use of videos picturing current states of environmental situations will be used to increase immersion while players will, in the game, make economically or environmentally driven choices. A teamed second player will be constrained by the first player’s decisions, and the first player needs to consider consequences for his/her follower. The imagery of environmental impact will also be used to track psycho-physiological responses to the real world via in-game cinematics by the use of wearables. The insights gained will be discussed, post statistical analyses, of psychological implications with all players in a town-hall meeting.

Background: In real-world scenarios humans perceive the world contextually, relying on previous information to modify their responses. During interactions with a machine, missing contexts may decrease the accuracy of judgements. In the realm of human-computer interactions (HCI), relatively easy tasks as controls may not be relevant. To evaluate the impact of stress we increased the cortisol level by the safe but reliable procedure Cold Pressor Task. We used five stimuli represented by facial expressions: ‘neutral’, ‘laughter’, ‘fear’, ‘pain’, and ‘pleasure’. Aim: We intend to find out how the responses to stimuli are altered by stress and statistically quantify the BVP (Blood Volume Pulse) signals. Materials: 27 raters rated these five stimuli presented by 5 actors and 5 actresses, while BVP was being registered. Methods: Each physiological response was a six-second time series after the rater rated the stimulus. A nontrivial model includes lag dependencies on either previous states or previous noise. The simplest models would be ARMA( p,q) models with to-be-determined parameters φ_1 , … φ_p and θ_1 , … θ_q . Inferences: In this study, we find that the wearables’ sampling for six seconds cannot separate signal from noise significantly. Only one response was found to be significantly affected by the condition of stress: the perception of fear.

Uvod: Infekce cevnich protez v aortoilicke oblasti (abdominalni VGI) je jednou z nejzavažnějsich komplikaci v cevni chirurgii. Řeseni je zatiženo vysokou mirou mortality a morbidity. V 2020 vydala the European Society for Vascular Surgery (ESVS) doporuceni pro diagnostiku a terapii infekci vaskularnich graftů a endograftů. I ve světle těchto doporuceni jsme se rozhodli retrospektivně prozkoumat pacienty s abdominalni VGI, ktere jsme řesili na nasem pracovisti.Metody: Retrospektivni observacni studie pacientů s abdominalni VGI, kteři byli řeseni na nasem pracovisti v obdobi 2011−2019 (za 9 let). Primarnim cilem bylo urcit procento infekci cevnich chirurgických rekonstrukci v aortoilicke oblasti provedených v obdobi 2011−2019 a zhodnotit mortalitu u pacientů operovaných pro tuto komplikaci. Sekundarnim cilem bylo zhodnotit uspěsnost a rizika různých typů výkonů.Výsledky: V uvedenem obdobi jsme provedli 363 chirurgických rekonstrukci v aortoilicke oblasti. Ve stejnem obdobi jsme řesili celkem 15 pacientů s abdominalni VGI, z nichž byla primarni aortalni rekonstrukce větsinou (11×) provedena před rokem 2011. U naseho souboru rekonstrukci z let 2011−2019 byla abdominalni VGI zaznamenana pouze ve 4 připadech, tedy u 1,1 %. V souboru 15 řesených abdominalnich VGI převažovali muži (14×). Průměrný věk v době původni rekonstrukce byl 61 let. Větsinou slo o ischemickou chorobu dolnich koncetin (14×). Vždy slo o infekci aortobifemoralniho (1× aortofemoralniho) bypassu. Vždy slo o pozdni infekci s průměrným intervalem 61 měsiců od původni rekonstrukce (15−180 měsiců). Casna mortalita v tomto souboru byla 27 % (4 pacienti). Celkova mortalita byla 40 %. Po primarnim řeseni infekce jsme zaznamenali 33 % reinfekci...