Bayesian Networks

En el enfoque bayesiano de la teoría de la probabilidad, una probabilidad cuantifica un grado de creencia asociado a un único ensayo o evento individual, sin ninguna conexión a priori con frecuencias límite obtenidas mediante la repetición indefinida de experimentos. En este trabajo mostramos que, a pesar de estar prescritas por una ley fundamental de la naturaleza, las probabilidades asociadas a sistemas cuánticos individuales pueden comprenderse de manera consistente dentro del marco conceptual bayesiano. Sostenemos que la diferencia entre las probabilidades clásicas y las probabilidades cuánticas no reside en su definición matemática ni en su interpretación probabilística básica, sino en la naturaleza de la información que dichas probabilidades codifican acerca de los sistemas físicos. En el mundo clásico, la información máxima acerca de un sistema físico es también completa, en el sentido de que permite asignar respuestas definidas a todas las preguntas físicamente significativas que puedan formularse sobre dicho sistema. En contraste, en el mundo cuántico, la información máxima no es completa y, además, no puede completarse mediante ninguna descripción física más detallada compatible con la teoría cuántica. Esta característica constituye una diferencia conceptual fundamental entre la física clásica y la mecánica cuántica, y se encuentra estrechamente relacionada con principios tales como la incompatibilidad de observables y las limitaciones impuestas por la estructura matemática de los estados cuánticos. Utilizando esta distinción, demostramos que toda asignación bayesiana de probabilidades en mecánica cuántica debe adoptar necesariamente la forma de la regla de probabilidad cuántica. Mostramos además que la posesión de información máxima sobre un sistema cuántico conduce a una asignación única de estado cuántico, y que la teoría cuántica establece una relación más fuerte entre probabilidad y frecuencia observada que la que puede justificarse dentro de un marco puramente clásico. Finalmente, presentamos una formulación bayesiana de la tomografía de estados cuánticos, entendida como el procedimiento inferencial mediante el cual se actualiza el conocimiento sobre un sistema cuántico a partir de los resultados experimentales obtenidos en un conjunto de mediciones.

This dissertation responds to the need for integration between researchers and decision-makers who are dealing with complex social-ecological system sustainability and decision-making challenges. To this end, we propose a new approach, called Bayesian Participatory-based Decision Analysis (BPDA), which makes use of graphical causal maps and Bayesian networks to facilitate integration at the appropriate scales and levels of descriptions. The BPDA approach is not a predictive approach, but rather, caters for a wide range of future scenarios in anticipation of the need to adapt to unforeseeable changes as they occur. We argue that the graphical causal models and Bayesian networks constitute an evolutionary, adaptive formalism for integrating research and decision-making for sustainable development. The approach was implemented in a number of different interdisciplinary case studies that were concerned with social-ecological system scale challenges and problems, culminating in a study where the approach was implemented with decisionmakers in Government. This dissertation introduces the BPDA approach, and shows how the approach helps identify critical cross-scale and cross-sector linkages and sensitivities, and addresses critical requirements for understanding system resilience and adaptive capacity.

Socioeconomic and ecological systems exhibit complex, interdependent behaviour which is often difficult to model and understand. This is due to the complex reorganisation of key sub-system processes involving nonlinear, cross-scale and cross-sector interactions in real time. Hence, predictive models of complex social-ecological systems are often subject to large uncertainties. We propose an approach for evaluating land-use adaptations for biofuel production, using Bayesian networks and integrating research on the food, water and energy sectors. The approach is intended to facilitate interdisciplinary consideration of cross-scale and intersector dependencies. We applied this approach to 2 examples of land-use strategies and show how the resilience of a strategy that meets the new South African national biofuel production target can be assessed in relation to climate change. Cross-disciplinary consideration of variables may be enhanced through the sensitivity analysis enabled by Bayesian networks, which is used to conceptualise the conditional causal dependencies between subsystem variables. We formulate and run a national scale South African model which links the impacts of projected climate change effects in southern Africa to irrigated agriculture, water storage planning and biofuel production. We demonstrate how the approach can be used to evaluate land-use changes in different projected climate change scenarios and land-use combinations, and how conflicting demands between water, food and biofuel energy sources may be preliminarily identified and assessed in an integrated probabilistic framework. Evaluating this problem in the context of climate change and water-related limits to growth enables research to support integrated analysis and planning for biofuel production and development. KEY WORDS: Bayesian networks • Climate change • Biofuels • Land-use change • Resilience • Cross-sector analysis Resale or republication not permitted without written consent of the publisher Contribution to CR Special 20 'Integrating analysis of regional climate change and response options'

We exploit the use of a controller area network (CAN-bus) to monitor sensors on the buses of local public transportation in a big European city. The aim is to advise fleet managers and policymakers on how to reduce fuel consumption so that air pollution is controlled and public services are improved. We deploy heuristic algorithms and exhaustive ones to generate Bayesian networks among the monitored variables. The aim is to describe the relevant relationships between the variables, to discover and confirm the possible cause–effect relationships, to predict the fuel consumption dependent on the contextual conditions of traffic, and to enable an intervention analysis to be conducted on the variables so that our goals are achieved. We propose a validation technique using Bayesian networks based on Granger causality: it relies upon observations of the time series formed by successive values of the variables in time. We use the same method based on Granger causality to rank the Bayesian ...

Este trabalho apresenta o processo de construção de um modelo de inferência de emoções que um aluno sente em relação a outros alunos durante interação síncrona em um contexto de jogo colaborativo de aprendizagem. A inferência de emoções está psicologicamente fundamentada na abordagem da avaliação cognitiva e foram investigadas relações entre objetivos e normas comportamentais do aluno e aspectos de sua personalidade. Especificamente, foram empregados o modelo OCC de emoções e o modelo Big-Five (Cinco Grandes Fatores) de traços de personalidade para a fundamentação teórica da modelagem. O modelo afetivo representa a vergonha e orgulho apresentados pelo aluno em resposta à avaliação cognitiva de suas próprias ações; e a reprovação e admiração apresentadas pelo aluno em resposta a ações de seu parceiro de jogo, a partir da avaliação do comportamento observável dos parceiros representado por suas interações no jogo colaborativo, em relação a normas comportamentais do aluno. A fim de sup...

Güncellenmiş 3. baskısını yapan kitap, uzun yıllar üniversitelerde "Benzetim ve modelleme", "Sistem benzetimi" ve "Simülasyon" derslerini anlatan yazarın notlarından, ders anlatımı esnasında öğrencilerinden gelen soru ve cevaplardan, öğrenci proje örneklerinden ve mesleki tecrübelerinden oluşmaktadır. Kitap ilgili derslere uygun olarak hazırlanmıştır. Bu kitap, üretim ve yönetim sistemlerine yönelik Arena ve ProModel yazılımları ile desteklenerek modelleme yapmak isteyen; öğrenci ve iş dünyası, akademisyen ve araştırmacılara yararlı olacak şekilde hazırlanmıştır. Okuyucular bu kitap ile sistem davranışını anlayacak, farklı girdi parametreleri ile üretim ve hizmet sistemlerinin çeşitli senaryolar altında performans tahminini hesaplayabilecektir.

The present era of large-scale foundation models is defined by a philosophical and practical tension: the successful, scalable application of frequentist Maximum Likelihood Estimation (MLE) and Maximum A Posteriori (MAP) optimization, which bypasses the theoretically superior but computationally intractable Bayesian ideal of full posterior marginalisation. While MLE-driven models achieve remarkable empirical results, they inherently lack the capacity to quantify epistemic uncertainty, leading to vulnerabilities like overconfidence and brittle generalization when encountering out-of-distribution data. This report examines the foundational advantages of Bayesian inference—automatic regularisation, model averaging, and calibrated probabilities—and the "curse of dimensionality" that necessitated the pivot to point-estimation scaling laws. The analysis further reveals a profound mathematical equivalence between Bayesian neural networks and quantum physics, specifically the Feynman path integral, suggesting that training a neural network is analogous to finding the ground state of a quantum many-body system. Finally, the paper evaluates cutting-edge solutions, including hybrid algorithmic frameworks like PoLAR-VBLL that surgically reintroduce uncertainty into Large Language Models, and argues that the ultimate resolution lies in a hardware paradigm shift toward stochastic neuromorphic memristors and quantum computing, which are physically engineered to perform probabilistic inference.

The dominant paradigm in artificial intelligence rests on an implicit meta-layer assumption: that the world is best modelled by learning joint probability distributions over observed variables. We argue this assumption is ontologically inverted. The physical, biological, and social world is governed by causal mechanisms that generate observations-observations are not primary features of reality but downstream projections of underlying causal structure. We formalise this position as the Generative Ontology and derive from it the Causal World Model Thesis: that world models built on causal foundations generalise from fewer observations, support intervention reasoning natively, handle feedback dynamics faithfully, require substantially less compute at inference, and produce discoveries that are interpretable and transferable across distributions. We contrast this with the statistical approach-including Joint Embedding Predictive Architectures (JEPA)-and argue that statistical world models are attempting to reconstruct generative mechanisms by working backwards from their outputs, which is both computationally harder and mathematically underdetermined. We further argue that causal models are not only superior reasoning engines for known domains but represent the correct architectural foundation for discovery in a causal world-composing known mechanisms to identify new configurations rather than inducing patterns from observations. We ground these arguments in causal inference theory, the history of natural science, and empirical observations from deployed causal AI systems across financial crime, clinical diagnosis, and macroeconomic domains.

Behavior of humpback whales was observed during the reproductive period off the northern coast of the state of Bahia (NB, n = 378 groups) and at the Abrolhos Bank (AB, n = 919) to compare patterns and group composition between the two locations. Alone individuals and dyads were most often encountered in both areas, although mother-calf pairs were more common in AB. While these two regions comprise distinct concentrations of humpback whales, with intrinsic environmental differences, behavior patterns were quite similar. The only behavioral differences found where for "tail…

Unlike conventional neural networks trained via gradient descent on static datasets, this architecture implements online predictive coding with no separation between training and inference. All learning occurs via local Hebbian and STDP rules during single-pass streaming. The system is self-modifying with formally verified, rollback-enabled safety bounds for mutations. Eighteen layers span spiking neural dynamics, PAD affective lattice, four￾store complementary memory, AGI planning, a full closed embodiment loop over RTSP
vision/audio/MAVLink drones/ROS robots, 73-state network reconnaissance, multi-chain blockchain with
MEV and flash loans, COBOL/TN3270 mainframe access and code-command capability (as well as construction of novel code for these systems), ICS/SCADA industrial command and write access, a 60-
channel encrypted distributed protocol, and a 4D spatiotemporal imagination pipeline seeded by live affective state, and countless other features detailed herein.

The global development policy has taken a significant issue in sustainable consumption, with a particular focus on the United Nations Sustainable Development Goals (SDGs). SDG 12: Responsible Consumption and Production focuses on responsible utilization of resources, sustainable consumption and minimization of food losses. Food consumption behaviour of households is therefore vital in the formulation of policies that can be sustainable and equitable. In this paper, the determinants of household food spending on urban households in Mumbai are studied in the period between 2022 and 2025 using the Engel curve analysis and regression models and dwells down to the implication on the sustainable consumption patterns.
The socio-economic variables studied in the research are household-level, such as income, age structure, gender structure, occupation, education, and household size. The key determinants of food expenditure are identified with the help of the multiple regression analysis, and the relationship between income and food consumption is examined with the help of the Engel curve estimation. The regression models show decent explanatory power with the R 2 of 0.352 to 0.418, which implies that a significant part of the food expenditure variation is attributed to income and demographic factors.
The results indicate that the most predictable and statistically significant predictor of food spending in all years is the household income. The estimation of the Engel curves proves that there is a positive correlation existing between food spending and income which shows that food is a normal good. Nonetheless, the decreasing marginal spending on food as income rises indicated in the concave form of the curves is in line with the Law of Engel. The response to income in the year 2023 is the best, indicating the recovery of consumption after the pandemic, and the balance expenditures in the 20242025 year.
In terms of sustainability, the research points out that demographic characteristics which include household size, education, and gender balance are factors that impact more on consumption behaviour, as time goes by. These are socio-economic determinants that play a vital role in the development of responsible consumption and allocation of resources at the household level. The results are relevant to SDG-12 as they illustrate how the city food consumption patterns are influenced by income dynamics and demographic peculiarities. The findings are valuable to policymakers who want to facilitate sustainable consumption, enhance food security and come up with welfare schemes that can compel responsible household expenditure habits within urban economies.
Keywords: Sustainable Consumption, Engel Curve, Food Spending, SDG 12, Household Consumption

L'accident de Fukushima a provoque une prise de conscience sans precedent concernant l'importance des risques naturels a l'echelle internationale. En mettant en cause plusieurs aspects institutionnels, et notamment les liens de dependance entre l'autorite de surete nucleaire et le ministere de l'industrie japonais, les analyses de l'accident posent egalement la question de la gouvernance et de la regulation des risques nucleaires avec une acuite nouvelle (Delamotte 2013 ; IAEA 2015). Le projet de recherche pluri-institutionnel AGORAS, lance en 2013, s'interesse ainsi a la gestion interorganisationnelle des risques et des accidents nucleaires. Ce projet vise une meilleure comprehension des regimes de regulation (Hood, Rothstein et al. 2001) des risques nucleaires et de leur dynamique d'evolution. La presente communication, qui s'inscrit dans cette action de recherche, s'interesse a plusieurs instruments de regulation des risques. Ceux-ci sont e...

Groundwater is the only source of fresh water in the Gaza Strip. However, it is severely polluted and requires immediate effort to improve its quality and increase its usable quantity. Intensive exploitation of groundwater in the Gaza Strip over the past 40 years has disturbed the natural equilibrium between fresh and saline water, and has resulted in increased salinity in most areas. Salinization in the coastal aquifer may be caused by a single process or a combination of different processes, including seawater intrusion, upconing of brines from the deeper parts of the aquifer, flow of saline water from the adjacent Eocene aquifer, return flow from irrigation water, and leakage of wastewater. Each of these sources is characterized by a distinguishable chemistry and well known isotopic ratios. In this paper Na/Cl, SO 4 /Cl, Br/Cl, Ca/(HCO 3 CSO 4 ), and Mg/Ca ionic ratios were used to distinguish different salinization sources. d 11 B and 87 Sr/ 86 Sr isotopic composition were also included in the model to study their importance in this monitoring task. The task of monitoring and the associated decision making process are characterized by a high degree of uncertainty with respect to input data and accuracy of models. For this reason, probabilistic expert systems, and more specifically, Bayesian belief networks (BBNs) are used to identify salinization origins. The BBN model incorporates the theoretical background of salinity sources, area-specific monitoring data that are characteristically incomplete in their coverage, expert judgment, and common sense reasoning to produce a geographic distribution for the most probable sources of salinization. The model is also designed to show areas where additional data on chemical and isotopic parameters are needed.

This paper conducts a comparative analysis of portfolio optimization methods, focusing on Bayesian approaches, applied to U.S. AI-related stocks (2020-2025). While the classical Markowitz model relies on fixed estimates of return and risk, the Bayesian framework incorporates parameter uncertainty through priors on expected returns, soft portfolio concentration constraints, and weights parameterized via Dirichlet or softmax transformations. Posterior inference is conducted using Hamiltonian Monte Carlo with the No-U-Turn Sampler (NUTS), allowing more adaptive and probabilistically informed decision-making. Portfolio performance is evaluated using risk-adjusted returns, measured by the Sharpe ratio, and supplemented with conditional volatility and beta dynamics via the conditional CAPM and DCC-GARCH framework. In the reported experiment, the Markowitz model achieved the highest Sharpe ratio (0.049 in-sample; 0.089 out-of-sample), but this result is limited by the narrow stock universe, daily frequency, zero risk-free rate assumption, and the exclusion of transaction costs. Advanced Bayesian models showed improved risk-adjusted performance relative to early Bayesian specifications, reaching up to 0.038 in-sample and 0.084 out-of-sample Sharpe ratios, while simultaneously reducing conditional volatility and market beta.

This thesis presents a comprehensive mathematical framework for determining the probability that we are living in a computer simulation. Building upon Nick Bostrom's foundational simulation argument, we develop a novel integrated approach that combines information-theoretic bounds, quantum mechanical evidence, computational complexity analysis, and Bayesian inference to provide a quantitative assessment of the simulation hypothesis. Our framework introduces several key innovations: (1) a discreteness metric based on quantum measurement precision, (2) computational feasibility analysis across different scales of reality, (3) observer effect quantification as evidence for simulation-like behavior, and (4) a unified Bayesian model that integrates multiple lines of evidence.

Through comprehensive analysis of four distinct scenarios (Conservative, Moderate, Optimistic, and Quantum-Focused), we find simulation probabilities ranging from 0.43 to 0.71, with confidence intervals providing statistical rigor to our estimates. Our empirical validation against known physical constants, quantum mechanical phenomena, and computational systems demonstrates strong support for the simulation hypothesis, with an overall validation score of 11.50. The framework reveals that while simulating systems up to solar system scale appears computationally feasible, simulating the entire observable universe exceeds current theoretical limits by 19 orders of magnitude.

Key findings include: (1) quantum discreteness at the Planck scale strongly suggests digital rather than continuous physics, (2) the observer effect in quantum mechanics aligns with computational optimization principles, (3) Bell inequality violations can be elegantly explained through simulation constraints, and (4) the fine-tuning of physical constants is consistent with computational parameter selection. This work provides the first rigorous mathematical framework for simulation detection and establishes a foundation for future empirical testing of the simulation hypothesis.

Both intensional and extensional background knowledge have previously been used in inductive problems to complement the training set used for a task. In this research, we propose to explore the usefulness, for inductive learning, of a new kind of intensional background knowledge: the inter-relationships or conditional probability distributions between subsets of attributes. Such information could be mined from publicly available knowledge sources but including only some of the attributes involved in the inductive task at hand. The purpose of our work is to show how this information can be useful in inductive tasks, and under what circumstances. We will consider injection of background knowledge into Bayesian Networks and explore its effectiveness on training sets of different sizes. We show that this additional knowledge not only improves the estimate of classification accuracy it also reduces the variance in the accuracy of the model.

This paper proposes two parallel variants of an Estimation of Distribution Algorithm (EDA) that represents the probability distribution by means of a single connected graphical model based on a polytree structure. The main goal is to design a new and more ef cient EDA. Our algorithm is based on the master/slave model that allows to perform the estimation of the probability distribution (the most time-consuming phase in EDAs) in a parallel way. The aim of our experimental studies is manifold. Firstly, we show that our parallel versions achieve a notable reduction of the total execution time with respect to existing algorithms. Secondly, we study the behavior of the algorithm from the numerical point of view, analyzing the different versions. Finally, our methods are evaluated over three interconnection networks (Fast Ethernet, Gigabit Ethernet, and Myrinet) and a study on the in uence of the parallel platform in the communication is performed.

Multiple stressors are an increasing concern in the management and conservation of ecosystems, and have been identified as a key gap in research. Coral reefs are one example of an ecosystem where management of local stressors may be a way of mitigating or delaying the effects of climate change. Predicting how multiple stressors interact, particularly in a spatially explicit fashion, is a difficult challenge. Here we use a combination of an expert-elicited Bayesian network (BN) and spatial environmental data to examine how hypothetical scenarios of climate change and local management would result in different outcomes for coral reefs on the Great Barrier Reef (GBR), Australia. Parameterizing our BN using the mean responses from our experts resulted in predictions of limited efficacy of local management in combating the effects of climate change. However, there was considerable variability in expert responses and uncertainty was high. Many reefs within the central GBR appear to be at ...

As is well known, the implementation of science-based achievements of different scientific disciplines for the purpose of criminal and judicial process in general, is one of the more prevalent tendencies of contemporary justice. In criminal proceedings, implementation of science-based achievements, in modern times, is expressed in two areas. First, is the area of obtaining evidence, and second, the area of interpretation and evaluation of probative value. In the area of obtaining evidence, scientific and technological development has reached such a level that they not only significantly facilitates the search for evidence thanks to sophisticated forensic procedures, but it also makes it possible to actively generate evidence, as is the case of technical recordings in special investigative techniques, for example. For the purpose of this paper, we are

En este taller se propone una herramienta para el análisis de los problemas de probabilidad condicional y de sus procesos de resolución, el grafo trinomial. El taller consta de cuatro fases: introducción al lenguaje del grafo, traducción de un problema al grafo y realización de lecturas analíticas, del grafo a la formulación de problemas y finalmente análisis de resoluciones mediante grafos. Palabras clave: Resolución de problemas, probabilidad condicional, didáctica de la probabilidad, formación de profesores de matemáticas. 1 Este trabajo forma parte del Proyecto EDU2008-03140/EDU financiado por el Ministerio de Ciencia y Tecnología (España).

Closed Circuit Television systems in shopping malls could be used to monitor the shopping behavior of people. From the tracked path, features can be extracted such as the relation with the shopping area, the orientation of the head, speed of walking and direction, pauses which are supposed to be related to the interest of the shopper. Once the interest has been detected the next step is to assess the shopper's positive or negative appreciation to the focused products by analyzing the (non-)verbal behavior of the shopper. Ultimately the system goal is to assess the opportunities for selling, by detecting if a customer needs support. In this paper we present our methodology towards developing such a system consisting of participating observation, designing shopping behavioral models, assessing the associated features and analyzing the underlying technology. In order to validate our observations we made recordings in our shop lab. Next we describe the used tracking technology and the results from experiments.

This volume provides a mathematically rigorous examination of advanced Bayesian models applied to private equity (PE) risk management under conditions of severe structural uncertainty. Recognizing the limitations of traditional frequentist frameworks like Maximum Likelihood Estimation (MLE) in illiquid, data-scarce private markets, this report demonstrates how Bayesian methods provide fully calibrated measures of uncertainty by representing all unknown parameters as probability distributions. The analysis covers algebraic and geometric interpretations of Bayesian inference and evaluates specific applications, including: Bayesian desmoothing to restate the true volatility and correlation of appraisal-based assets; Bayesian Hierarchical Models (BHMs) for modeling the extreme right-skewness and heavy-tailed payoffs of fund performance and cash flow dynamics; Markov-Switching Bayesian Vector Autoregressions (MS-BVAR) for dynamically adapting to abrupt macroeconomic regime shifts and parsing external shocks in volatile emerging markets like South Asia; and Bayesian Belief Networks (BBNs) for quantifying qualitative risks such as political and regulatory threats. The report concludes by exploring the frontier of Quantum Bayesian Machine Learning (QBML), which leverages Quantum Amplitude Estimation (QAE) and optimization algorithms like QAOA and VQE to overcome classical computational bottlenecks and enable real-time, dynamic portfolio optimization. Ultimately, the synthesis demonstrates that transitioning to full posterior predictive distributions allows for a more resilient, dynamic, and uncertainty-aware approach to institutional capital allocation. Mamun, SM (2026) …………………………………………………...1 Advanced Applications of Bayesian Models for Private Equity Risk Management Under Structural Uncertainties: (Volume II)

This paper argues that modern artificial intelligence systems are fundamentally misaligned with real-world domains because they optimize symmetric predictive objectives such as expected loss, while consequential environments are governed by asymmetric consequences and discontinuous failure boundaries. The work proposes an alternative framework in which expectations rather than predictions serve as the primitive of intelligence. Expectations encode commitments about what the environment will permit within specified tolerances, and violations provide the primary learning signal. The framework introduces mechanisms for violation accumulation, structural regime change detection through co-violation topology, and planning under explicit viability thresholds. Large language models are integrated as bounded proposal engines whose outputs must pass constraint-based admission gates. The result is a constraint-governed architecture designed to maintain system viability and reliability in enforcement-governed domains.

This work proposes a novel approach for solving abductive reasoning problems in Bayesian networks involving fuzzy parameters and extra constraints. The proposed method formulates abduction problems using nonlinear programming. To maximize the sum of the fuzzy membership functions subjected to various constraints, such as boundary, dependency and disjunctive conditions, unknown node belief propagation is completed. The model developed here can be built on any exact propagation methods, including clustering, joint tree decomposition, etc.

Büyük ölçekli sanayi ve enerji yatırımlarında alınan kararlar, çoğu zaman yalnızca teknik yeterlilik ve maliyet hesabıyla şekillenmez. Kurumsal çıkarlar, danışman etkileri, algı yönetimi ve görünmeyen güç dengeleri, sonuç üzerinde belirleyici olabilir. Türkiye'de 1990'lı yıllarda yaşanan ve bir enerji ekipmanı tedarik sürecinin kaybedilmesiyle sonuçlanan örnek olay, bu gerçeği tüm açıklığıyla ortaya koymaktadır.

The traditional approach to the study of human factors in the maritime field involves the analysis of accidents without considering human factors reliability analysis. The main approach being use to analyze human errors are statistical approach and probability theory approach. Another suitable approach to the study of human factors in the maritime industry is the quasi-experimental field study where variations in performance (for example attention) can be observed as a function of natural variations in performance shaping factors. This paper analyzes result of modeling human error and human reliability emanating from the use of technology on board ship navigation in coastal water area by using qualitative and quantitative tools. Accident reports from marine department are used as empirical material for quantitative analysis. The literature on safety is based on common themes of accidents, the influence of human error resulting from technology usage design, accident report from MAIB and interventions information are use for qualitative assessment. Human reliability assessment involves analysis of accident in waterways, emanating from human-technology factors interface. This paper report an enhancement requirement of the methodological issues with previous research study, monitoring and deduce recommendations for technology modification of the human factors necessary to improve maritime safety performance. The result presented can contribute to rule making, and safety management leading for development of guideline and standards for human reliability risk management for ship navigating within inland and coastal waters.

Clustering and correlation analysis techniques have become popular tools for the analysis of data produced by metabolomics experiments. The results obtained from these approaches provide an overview of the interactions between objects of interest. Often in these experiments, one is more interested in information about the nature of these relationships, e.g., cause-effect relationships, than in the actual strength of the interactions. Finding such relationships is of crucial importance as most biological processes can only be understood in this way. Bayesian networks allow representation of these cause-effect relationships among variables of interest in terms of whether and how they influence each other given that a third, possibly empty, group of variables is known. This technique also allows the incorporation of prior knowledge as established from the literature or from biologists. The representation as a directed graph of these relationship is highly intuitive and helps to understand these processes. This paper describes how constraint-based Bayesian networks can be applied to metabolomics data and can be used to uncover the important pathways which play a significant role in the ripening of fresh tomatoes. We also show here how this methods of reconstructing pathways is intuitive and performs better than classical techniques. Methods for learning Bayesian network models are powerful tools for the analysis of data of the magnitude as generated by metabolomics experiments. It allows one to model cause-effect relationships and helps in understanding the underlying processes.

Socio-ecological systems are recognized as complex adaptive systems whose multiple interactions might change as a response to external or internal changes. Due to its complexity, the behavior of the system is often uncertain. Bayesian networks provide a sound approach for handling complex domains endowed with uncertainty. The aim of this paper is to analyze the impact of the Bayesian network structure on the uncertainty of the model, expressed as the Shannon entropy. In particular, three strategies for model structure have been followed: naive Bayes (NB), tree augmented network (TAN) and network with unrestricted structure (GSS). Using these network structures, two experiments are carried out: (1) the impact of the Bayesian network structure on the entropy of the model is assessed and (2) the entropy of the posterior distribution of the class variable obtained from the different structures is compared. The results show that GSS constantly outperforms both NB and TAN when it comes to...

Continual learning systems must remain plastic enough to adapt to evolving data while retaining previously acquired knowledge. Bayesian methods offer a natural framework for continual learning, but standard sequential Bayesian inference accumulates evidence indefinitely and can become over-confident, hindering adaptation to distributional shift. In this work, we develop a general theory and practical algorithms for Bayesian forgetting: explicit mechanisms that allow Bayesian models to selectively discard stale information while preserving relevant structure. We formalize Bayesian forgetting as an operator on posterior distributions, characterize its effect on effective sample size, and show how it recovers and unifies classical discounting schemes and sliding-window posteriors. We derive efficient approximate inference algorithms for Bayesian neural networks based on variational inference and provide idealized, exponential-family-based regret heuristics suggesting when Bayesian forgetting can track drifting targets while controlling dynamic loss. Simulation-based experiments in Google Colab on regression and toy memory benchmarks with controlled concept drift illustrate the resulting plasticity-stability trade-off and qualitative links to inverse-Bayesian phenomena.

Probabilistic topic models are a popular tool for the unsupervised analysis of text, providing both a predictive model of future text and a latent topic representation of the corpus. Recent studies have found that while there are suggestive connections between topic models and the way humans interpret data, these two often disagree. In this paper, we explore this disagreement from the perspective of the learning process rather than the output. We present a novel task, tag-and-cluster, which asks subjects to simultaneously annotate documents and cluster those annotations. We use these annotations as a novel approach for constructing a topic model, grounded in human interpretations of documents. We demonstrate that these topic models have features which distinguish them from traditional topic models.

Closed Circuit Television systems in shopping malls could be used to monitor the shopping behavior of people. From the tracked path, features can be extracted such as the relation with the shopping area, the orientation of the head, speed of walking and direction, pauses which are supposed to be related to the interest of the shopper. Once the interest has been detected the next step is to assess the shopper's positive or negative appreciation to the focused products by analyzing the (non-)verbal behavior of the shopper. Ultimately the system goal is to assess the opportunities for selling, by detecting if a customer needs support. In this paper we present our methodology towards developing such a system consisting of participating observation, designing shopping behavioral models, assessing the associated features and analyzing the underlying technology. In order to validate our observations we made recordings in our shop lab. Next we describe the used tracking technology and the results from experiments.

Özet. İlişkisel veritabanı yönetim sistemleri (İVYS) bankacılık, telekom ve benzeri birçok alandaki kritik yazılımların işleyişinde önemli bir rol oynamaktadır. Bu sistemlerde oluşan hatalar ve kesintilerin büyük maliyetleri olabilmektedir. Bu sebeple bu sistemlerin sürekli takibi ve hataların önceden tahmin edilebilmesi büyük önem taşımaktadır. Mevcut durumda sistemlerde oluşan teknik hatalar uzman personel tarafından takip edilmekte ve sorun oluştuktan sonra personel tarafından çözüm üretilmesi beklenmektedir. Bu da verimlilik ve zaman kaybına neden olmaktadır. Sorunlar genellikle kritik hale geldiğinde ortaya çıktığı için hem çözümün uygulanması zorlaşmakta hem de kullanıcıların iş süreçleri sorundan etkilenmektedir. Bu çalışmada belirtilen yöntemlerle veritabanı ve orta katman ürünlerini kullanan firmalardaki sistem altyapılarında meydana gelen teknik problemlerin, ortaya çıkmadan önce tahmin edilerek, düzeltici aksiyonların önerilmesini sağlayacak makine öğrenmesi tabanlı bir e...

This article presents a Bayesian implementation of a cumulative probit model to forecast the outcomes of the UEFA Champions League matches. The argument of the normal CDF involves a cut-off point, a home vs away playing effect and the difference in strength of the two competing teams. Team strength is assumed to follow a Gaussian distribution the expectation of which is expressed as a linear regression on an external rating of the team from eg. the UEFA Club Ranking (UEFACR) or the Football Club World Ranking (FCWR). Priors on these parameters are updated at the beginning of each season from their posterior distributions obtained at the end of the previous one. This allows making predictions of match results for each phase of the competition: group stage and knock-out. An application is presented for the 2013-2014 season. Adjustment based on the FCWR performs better than on UEFACR. Overall, using the former provides a net improvement of 24% and 23% in accuracy and Brier’s score over...

This paper develops a unified analytical framework for measuring political legitimacy across heterogeneous governance domains. Building on insights from constitutional political economy, social choice theory, and institutional analysis, the framework establishes consent-holding-the mapping from decision domains to those with authority over them-as a structural necessity of collective action. We formalize this intuition through five axioms and five theorems, demonstrating that legitimacy can be operationalized as stakes-weighted consent alignment α(d,t), while friction F(d,t) measures the deviation between outcomes and stakeholder preferences. The framework bridges normative democratic theory and empirical prediction, generating testable hypotheses about institutional stability. Historical validation examines suffrage expansion, abolition movements, labor rights, and contemporary platform governance, demonstrating how misalignment between stakes and voice generates observable instability. Unlike existing approaches that prescribe ideal institutions, this framework provides analytical tools for measuring legitimacy within any governance structure, enabling systematic comparison across democratic, technocratic, and algorithmic systems. Computational mechanism comparison via Bayesian learning dynamics across 1000 Monte Carlo runs demonstrates relative performance under adaptive agents: when preferences update based on observed policy outcomes, stakes-weighted DoCS achieves highest final alignment (α = 0.872) with lowest terminal friction (F = 1.5, 94.9% reduction from initial F = 30.3). This comparative advantage holds across static baseline (α = 0.627), learning dynamics (α = 0.872), and alternative temporal mechanisms, suggesting stakes-weighting produces superior initial matches that persist even when agents adapt to institutional performance. The framework's domain-specific approach resolves the apparent tension between consent and competence, showing both as complementary dimensions of institutional legitimacy. This framework is part of the Adversarial Systems Research program, which examines stability, alignment, and friction dynamics in complex systems where competing interests generate structural conflict.

Face recognition systems robust to major occlusions have wide applications ranging from consumer products with biometric features to surveillance and law enforcement applications. In unconstrained scenarios, faces are often subject to occlusions, apart from common variations such as pose, illumination, scale, orientation and so on. In this paper we propose a novel Bayesian oriented occlusion model inspired by psychophysical mechanisms to recognize faces prone to occlusions amidst other common variations. We have discovered and modeled similarity maps that exist in facial domains by means of Bayesian Networks. The proposed model is capable of efficiently learning and exploiting these maps from the facial domain. Hence it can tackle the occlusion uncertainty reasonably well. Improved recognition rates over state of the art techniques have been observed.

In this article we propose CPrefMiner, a mining technique for learning a Bayesian Preference Network (BPN) from a given sample of user choices. In our approach, user preferences are not static and may vary according to a multitude of user contexts. So, we name them Contextual Preferences. Contextual Preferences can be naturally expressed by a BPN. The method has been evaluated in a series of experiments executed on synthetic and real-world datasets and proved to be efficient to discover user contextual preferences.

The present paper is focusing on the connection between the Isomorphic groups and the Isomorphic graphs. We established a few results on the necessary and sufficient condition on the dimension of the graphs, which are isomorphic with respect to the isomorphic groups. We proved those results depending on the Isomorphic groups. We discussed the methods to find isomorphic graphs by using isomorphic groups. We derived few results on the isomorphism of graphs, by illustrating few examples.

This review critically examines the work of Cheng and Tong (2025) on overcoming the intractability barrier in infinite-dimensional non-parametric information geometry through an orthogonal decomposition of the tangent space. The paper introduces the Covariate Fisher Information Matrix (cFIM) and establishes foundational links to G-entropy, KL-divergence curvature, semi-parametric efficiency, and the Manifold Hypothesis. Salient points are summarized, highlighting theoretical breakthroughs and applications. Comparisons with prior research in information geometry, such as Amari's parametric framework and Pistone's nonparametric extensions, reveal contrasts in tractability and scope. Extensions are proposed, including applications to quantum information geometry, Bayesian nonparametrics, and high-dimensional causal inference. Systemic advantages over traditional approaches in information geometry are discussed, emphasizing enhanced explainability and computational feasibility. The review presents a coherent synthesis, underscoring the framework's potential to bridge abstract geometry with practical AI demands.

Nowadays, robotic platforms tend to be equipped with a conjugation of multi-modal artificial perception systems to navigate and interact with the surrounding environment and persons. The complexity and dynamic characteristics of those environments has led to the development of attention mechanisms to filter the sensory overload to sense, perceive and process only the relevant sensory data. This work presents Bayesian models related with the attentional mechanisms involved in blind manipulation of objects and related with the detection of borders, borders following and corner detection of object surfaces. This type of object features can be used as structural references of the manipulated object and be used in the next manipulation stages. The perception of those stimuli requires coordination between the attention, perception and action mechanisms in order to direct (attention) and promote (action) the contact between the touch sensors and the object, to maximize the acquisition of information (perception) and uncertainty reduction. The Bayesian model integrates information related with the surface curvature, fingers motion direction and surface texture. The magnitude of the variation of those object properties characterizes transition regions between between the object and the surrounding environment. The detection of these transitions allow the robotic hand to estimate an initial location of the borders of the object and start the board following using the robotic fingers and recognize the interception between borders -the corners. The statistical temporal relations between the action primitives required to explore and manipulate the object are based in an offline learning period. The action plan is inferred from this learned knowledge. The models of this work have been tested in a scenario that involved the blind manipulation of a towel and the detection of its corners. This scenario was implemented in ROS-Gazebo simulation environment, using the Shadow dexterous robotic hand model [1] equipped with the Shadow tactile sensor models attached to the fingertips of the robotic hand. [1] .