Probabilistic Model Research Articles

Bridging theory and data: A computational workflow for cultural evolution

Cultural evolution applies evolutionary concepts and tools to explain the change of culture over time. Despite advances in both theoretical and empirical methods, the connections between cultural evolutionary theory and evidence are often vague, limiting progress. Theoretical models influence empirical research but rarely guide data collection and analysis in logical and transparent ways. Theoretical models themselves are often too abstract to apply to specific empirical contexts and guide statistical inference. To help bridge this gap, we outline a quality-assurance computational workflow that starts from generative models of empirical phenomena and logically connects statistical estimates to both theory and real-world explanatory goals. We emphasize and demonstrate validation of the workflow using synthetic data. Using the interplay between conformity, migration, and cultural diversity as a case study, we present coded and repeatable examples of directed acyclic graphs, tailored agent-based simulations, a probabilistic transmission model for longitudinal data, and an approximate Bayesian computation model for cross-sectional data. We discuss the assumptions, opportunities, and pitfalls of different approaches to generative modeling and show how each can be used to improve data analysis depending on the structure of available data and the depth of theoretical understanding. Throughout, we highlight the significance of ethnography and of collecting basic cultural and demographic information about study populations and call for more emphasis on logical and theory-driven workflows as part of science reform.

Use of a Semi-Explicit Probabilistic Numerical Model for Concrete Cracking: From Static to Dynamic Loadings

In this paper, concrete cracking is investigated in dynamics through finite element modeling. A probabilistic semi-explicit model, previously developed and validated for static loading, is extended for dynamic loading. The model in statics is based on two material mechanical parameters: the tensile strength and the critical strain-energy release rate in mode I, GIC, of the Linear Elastic Fracture Mechanics (LEFM) theory. Concerning the dynamic aspects of the model, the tensile strength rate effect is modeled by an empirical dynamic-to-static strength ratio (Dynamic Increase Factor—DIF) and a similar formulation is proposed for GIC. The structural rate effect is naturally captured when mass and damping are included in the equation of motion. For static and dynamic loading, only macroscopic crack propagation is considered. Some numerical simulations in statics and dynamics are presented in the present paper. The main results related to this work can be summarized as follows: the dispersion of the numerical results related to the load–displacement curves decreases with the loading rate. The crack pattern considerably changes with loading rate (numerically and experimentally); the agreement between the experimental and numerical results (load–displacement curves and crack pattern) indicates the model is promising for engineering applications.

The probabilistic dependence of ship-induced waves is preserved spatially and temporally in the Savannah River (USA).

The rapid changes in the shipping fleet during the last decades has increased the ship-induced loads and, thus, their impact on infrastructures, margin protections and ecosystems. Primary waves have been pointed out as the cause of those impacts, with heights that can exceed 2m and periods around 2 minutes. Consequently, extensive literature can be found on their estimation mainly from a deterministic perspective with methods based on datasets limited to one location, making difficult their generalization. These studies propose either computationally expensive numerical models or empirical equations which often underestimate the extreme primary waves, hindering their use for design purposes. Moreover, a framework to allow the design of infrastructure under ship-wave attack based on probabilistic concepts such as return periods is still missing. In this study, a probabilistic model based on bivariate copulas is proposed to model the joint distribution of the primary wave height, the peak of the total energy flux, the ship length, the ship width, the relative velocity of the ship and the blockage factor. This model, a vine-copula, is developed and validated for four different deployments along the Savannah river (USA), with different locations and times. To do so, the model is quantified using part of the data in one deployment and validated using the rest of the data from this deployment and data of the other three. The vine-copula is validated from both a predictive performance point of view and with respect to the statistical properties. We prove that the probabilistic dependence of the data is preserved spatially and temporally in the Savannah river.

Modeling the velocity of evolving lineages and predicting dispersal patterns

Accurate estimation of the dispersal velocity or speed of evolving organisms is no mean feat. In fact, existing probabilistic models in phylogeography or spatial population genetics generally do not provide an adequate framework to define velocity in a relevant manner. For instance, the very concept of instantaneous speed simply does not exist under one of the most popular approaches that models the evolution of spatial coordinates as Brownian trajectories running along a phylogeny. Here, we introduce a family of models-the so-called Phylogenetic Integrated Velocity (PIV) models-that use Gaussian processes to explicitly model the velocity of evolving lineages instead of focusing on the fluctuation of spatial coordinates over time. We describe the properties of these models and show an increased accuracy of velocity estimates compared to previous approaches. Analyses of West Nile virus data in the United States indicate that PIV models provide sensible predictions of the dispersal of evolving pathogens at a one-year time horizon. These results demonstrate the feasibility and relevance of predictive phylogeography in monitoring epidemics in time and space.

Probabilistic projections of the Amery Ice Shelf catchment, Antarctica, under conditions of high ice-shelf basal melt

Abstract. Antarctica's Lambert Glacier drains about one-sixth of the ice from the East Antarctic Ice Sheet and is considered stable due to the strong buttressing provided by the Amery Ice Shelf. While previous projections of the sea-level contribution from this sector of the ice sheet have predicted significant mass loss only with near-complete removal of the ice shelf, the ocean warming necessary for this was deemed unlikely. Recent climate projections through 2300 indicate that sufficient ocean warming is a distinct possibility after 2100. This work explores the impact of parametric uncertainty on projections of the response of the Lambert–Amery system (hereafter “the Amery sector”) to abrupt ocean warming through Bayesian calibration of a perturbed-parameter ice-sheet model ensemble. We address the computational cost of uncertainty quantification for ice-sheet model projections via statistical emulation, which employs surrogate models for fast and inexpensive parameter space exploration while retaining critical features of the high-fidelity simulations. To this end, we build Gaussian process (GP) emulators from simulations of the Amery sector at a medium resolution (4–20 km mesh) using the Model for Prediction Across Scales (MPAS)-Albany Land Ice (MALI) model. We consider six input parameters that control basal friction, ice stiffness, calving, and ice-shelf basal melting. From these, we generate 200 perturbed input parameter initializations using space filling Sobol sampling. For our end-to-end probabilistic modeling workflow, we first train emulators on the simulation ensemble and then calibrate the input parameters using observations of the mass balance, grounding line movement, and calving front movement with priors assigned via expert knowledge. Next, we use MALI to project a subset of simulations to 2300 using ocean and atmosphere forcings from a climate model for both low- and high-greenhouse-gas-emission scenarios. From these simulation outputs, we build multivariate emulators by combining GP regression with principal component dimension reduction to emulate multivariate sea-level contribution time series data from the MALI simulations. We then use these emulators to propagate uncertainty from model input parameters to predictions of glacier mass loss through 2300, demonstrating that the calibrated posterior distributions have both greater mass loss and reduced variance compared to the uncalibrated prior distributions. Parametric uncertainty is large enough through about 2130 that the two projections under different emission scenarios are indistinguishable from one another. However, after rapid ocean warming in the first half of the 22nd century, the projections become statistically distinct within decades. Overall, this study demonstrates an efficient Bayesian calibration and uncertainty propagation workflow for ice-sheet model projections and identifies the potential for large sea-level rise contributions from the Amery sector of the Antarctic Ice Sheet after 2100 under high-greenhouse-gas-emission scenarios.

Bayesian-Neural-Network-Based Approach for Probabilistic Prediction of Building-Energy Demands

Reliable prediction of building-level energy demand is crucial for the building managers to optimize and regulate energy consumption. Conventional prediction models omit the uncertainties associated with demand over time; hence, they are mostly inaccurate and unreliable. In this study, a Bayesian neural network (BNN)-based probabilistic prediction model is proposed to tackle this challenge. By quantifying the uncertainty, BNNs provide probabilistic predictions that capture the variations in the energy demand. The proposed model is trained and evaluated on a subset of the building operations dataset of Lawrence Berkeley National Laboratory (LBNL), Berkeley, California, which includes diverse attributes related to climate and key building-performance indicators. We have performed thorough hyperparameter tuning and used fixed-horizon validation to evaluate trained models on various test data to assess generalization ability. To validate the results, quantile random forest (QRF) was used as a benchmark. This study compared BNN with LSTM, showing that BNN outperformed LSTM in uncertainty quantification.

A Hybrid Probabilistic Graph Based Community Clustering Model for Large Social Networking Link Prediction Data

Dynamic community clustering is essential for online social networking sites due to the high dimensionality and large data size. It aims to uncover social relationships among nodes and links within the network. However, traditional models often struggle with community structure detection because of the extensive computational time and memory required. Additionally, these models need contextual weighted node information to establish social networking feature relationships. To address these challenges, an advanced probabilistic weighted community detection framework has been developed for large-scale social network data. This framework uses a filter-based probabilistic model to eliminate sparse values and identify weighted community detection nodes for dynamic clustering analysis. Experimental results demonstrate that this filter-based probabilistic community detection framework outperforms others in terms of normalized mutual information, entropy, density, and runtime efficiency (measured in milliseconds).

A probabilistic model for API contract specification retrieval focusing on the openAPI standard

A probabilistic model for API contract specification retrieval focusing on the openAPI standard

Modelling construction performance variability for probabilistic time estimation of tunnelling projects

ABSTRACT Budget overrun and schedule delay of infrastructure projects result in the mismanagement of huge amounts of resources. Studies show that accounting for uncertainties in time and cost estimation of such projects can help the decision-makers in better understanding the involved risks. Several probabilistic models have been developed during the last two decades to facilitate such estimations. A common aspect of all these models is the subjective assessment of unit activities' duration, which affects the accuracy of the estimated time and cost significantly. Despite this, the mechanism governing the variability of unit activities' duration has seldom been studied. Thus this paper focuses on addressing this gap by using a unique set of data from a tunnel project. The variability in unit activities' duration (construction performance variability) is governed by three main components: typical performance variability, minor machinery delays and minor performance delays. Using these components, a novel method for modelling construction performance variability in probabilistic time estimation of tunnelling projects is proposed. The application of the proposed method is demonstrated through a case example with a discussion on its important aspects, advantages and limitations. The findings of this paper offer a resource to improve the accuracy of time estimation for tunnelling projects.

Heuristic satisficing inferential decision making in human and robot active perception

Inferential decision-making algorithms typically assume that an underlying probabilistic model of decision alternatives and outcomes may be learned a priori or online. Furthermore, when applied to robots in real-world settings they often perform unsatisfactorily or fail to accomplish the necessary tasks because this assumption is violated and/or because they experience unanticipated external pressures and constraints. Cognitive studies presented in this and other papers show that humans cope with complex and unknown settings by modulating between near-optimal and satisficing solutions, including heuristics, by leveraging information value of available environmental cues that are possibly redundant. Using the benchmark inferential decision problem known as “treasure hunt”, this paper develops a general approach for investigating and modeling active perception solutions under pressure. By simulating treasure hunt problems in virtual worlds, our approach learns generalizable strategies from high performers that, when applied to robots, allow them to modulate between optimal and heuristic solutions on the basis of external pressures and probabilistic models, if and when available. The result is a suite of active perception algorithms for camera-equipped robots that outperform treasure-hunt solutions obtained via cell decomposition, information roadmap, and information potential algorithms, in both high-fidelity numerical simulations and physical experiments. The effectiveness of the new active perception strategies is demonstrated under a broad range of unanticipated conditions that cause existing algorithms to fail to complete the search for treasures, such as unmodelled time constraints, resource constraints, and adverse weather (fog).

Physically based probabilistic prediction of regional rainfall-induced shallow landslides: from initiation to runout analysis

ABSTRACT In this study, a physically based method is proposed to estimate the exceedance probability of the intensity measure of the regional rainfall-induced shallow landslides. This method consists of a digital elevation model, an initiation model, a runout model and a probabilistic analysis model. In the suggested method, a study area is discretised into a grid of cells, and the stability of each cell can be evaluated through the initiation model. Then, the runout behaviour of the unstable cells can be simulated by the runout model. Finally, the exceedance probability of the intensity measures of each cell can be estimated through Monte Carlo simulation considering the uncertainties in the initiation model and runout model. The proposed method is applied to analyse the rainfall-induced landslides in a study area in the southwest of China, which can provide a reasonably accurate prediction of the impact zones of the landslides. The suggested method provides a useful tool to probabilistically predict the impacted area of the rainfall-induced landslides at a regional scale.

Development of a regionally consistent and fully probabilistic earthquake risk model for Central Asia

Abstract. A fully probabilistic earthquake risk model was developed for five countries in Central Asia, providing updated earthquake loss estimates with a higher level of detail on all components with respect to previous studies in the region. Additionally, a regionally consistent approach that, on the one hand, allows for direct comparisons at different disaggregation levels (e.g., country and oblast) was used. On the other hand, this updated earthquake risk assessment aims to facilitate initiating a policy dialogue regarding national and regional disaster risk financing and insurance applications. This earthquake risk model made use of a regional probabilistic seismic hazard analysis and a comprehensive exposure database on which different types of assets and sectors were included and for which two scenarios (the years 2020 and 2080) were modeled. For each type of asset, a unique vulnerability function was derived and later used for convolution with the hazard data, which allowed for the estimation of the loss exceedance curve, at different disaggregation levels, from where other risk metrics such as the average annual loss (AAL) and specific return period losses were obtained. The regional earthquake AAL for the five countries and for the 2020 exposure scenario has been estimated to be around USD 2 billion, with the Kyrgyz Republic and Tajikistan being the countries with the highest earthquake risk levels in the region. Besides the probabilistic earthquake risk results, as-if scenarios were modeled using a pseudo-deterministic approach to assess the human and economic losses for realistic and representative earthquakes for the main cities within earthquake-prone regions in the five countries of the study area.

A generalized multinomial probabilistic model for SARS-COV-2 infection prediction and public health intervention assessment in an indoor environment.

SARS-CoV-2 Omicron and its sub-lineages have become the predominant variants globally since early 2022. As of January 2023, over 664 million confirmed cases and over 6.7 million deaths had been reported globally. Current infection models are limited by the need for large datasets or calibration to specific contexts, making them difficult to apply to different settings. This study aims to develop a generalized multinomial probabilistic model of airborne infection to assist public health decision-makers in evaluating the effectiveness of public health interventions (PHIs) across a broad spectrum of scenarios. The proposed model systematically incorporates group characteristics, epidemiology, viral loads, social activities, environmental conditions, and PHIs. Assumptions about social distance and contact duration that estimate infectivity during short-term group gatherings have been made. The study is differentiated from earlier works on probabilistic infection modeling in the following ways: (1) predicting new cases arising from more than one infectious person in a gathering, (2) incorporating additional key infection factors, and (3) evaluating the effectiveness of multiple PHIs on SARS-CoV-2 infection simultaneously. Although the results show that limiting group size has an impact on infection, improving ventilation has a much greater positive health impact. The proposed model is versatile and can flexibly accommodate other scenarios or airborne diseases by modifying the parameters allowing new factors to be added.

Exploring the interaction of design variability and stochastic operational uncertainties in software-intensive systems through the lens of modeling

AbstractIn software-intensive systems, navigating the complexities that emerge from the interaction of design variability and stochastic operational uncertainties presents a daunting challenge. This paper delves into the dynamics between these two dimensions of uncertainty, offering novel insights about how modeling can contribute to the analysis of their combined impact upon system properties. By elevating the abstraction level at which probabilistic models are conceptualized, our approach enables an integrated analysis framework that considers both structural and quantitative dimensions of design spaces. Through the introduction of novel language constructs, our methodology facilitates the direct referencing of structural relationships within probabilistic behavioral specifications. Furthermore, the adoption of novel quantifiers in probabilistic temporal logic enables evaluating complex properties across diverse design variants, thereby streamlining the assessment of guarantees within the solution space. We demonstrate the feasibility of this approach on four case studies, showcasing its potential to offer comprehensive insights into the trade-offs and decision-making processes inherent in managing different types of structural design variability and operational uncertainties in software-intensive systems.

PAYOFF-BASED PROBABILISTIC INTERACTION MODEL ON THE EVOLUTION OF COOPERATION IN SPATIAL PUBLIC GOODS GAME

In the spatial public goods game (SPGG), punishment effectively promotes cooperation but often reduces the collective benefits of cooperators and punishers. In order to increase revenue, we propose a probabilistic interaction domain model considering both strategy type and payoff level. In this model, players are divided into two types, successful players with payoffs higher than the payoff threshold and failed players with payoffs lower than the payoff threshold. A successful player is less likely to change the interaction range than a failed player. Through extensive simulations, it is shown to verify that a higher payoff threshold leads to a more pronounced promotion effect on cooperation and corresponds to a higher cooperation return. Moreover, introducing dynamic interaction domain can rapidly remove defectors from the vicinity of cooperative players on regular lattices. Reducing the payoff gap between punishers and cooperators helps mitigate the system’s second-order free-riding problem. Additionally, through analysis of the critical parameters, it is revised that incorporating diversity in interaction structures substantively enhances cooperation level.

Photometric Redshift Estimation of Quasars by a Cross-modal Contrast Learning Method

Estimating photometric redshifts (photo-z) of quasars is crucial for measuring cosmic distances and monitoring cosmic evolution. While numerous point estimation methods have successfully determined photo-z, they often struggle with the inherently ill-posed nature of the problem and frequently overlook significant morphological features in the probability density functions (pdfs) of photo-z, such as calibration and sharpness. To address these challenges, we introduce a cross-modal contrastive learning probabilistic model that employs adversarial training, contrastive loss functions, and a mixture density network to estimate the pdf of photo-z. This method facilitates the conversion between multiband photometric data attributes, such as magnitude and color, and photometric image features, while extracting features invariant across modalities. We utilize the continuous ranked probability score (CRPS) and the probability integral transform (PIT) as metrics to assess the quality of the pdf. Our approach demonstrates robust performance across various survey bands, image qualities, and redshift distributions. Specifically, in a comprehensive data set from the Sloan Digital Sky Survey and the Wide-field Infrared Survey Explorer (WISE) survey, our probabilistic model achieved a CRPS of 0.1187. Additionally, in a combined data set from SkyMapper and WISE, it reached a CRPS of 0.0035. Our probabilistic model also produced well-calibrated PIT histograms for both data sets, indicating nearly uniform distributions. We further tested our approach in classification tasks within the SkyMapper data set. Despite the absence of u, v, and g bands, it effectively distinguished between quasars, galaxies, and stars with an accuracy of 98.96%. This versatile method can be extended to other scenarios, such as analyzing extended sources like galaxies, across different surveys and varying redshift distributions.

Optimizing embedded AI systems for autonomous driving: Challenges and solutions using bayesian networks

Abstract. Embedded Artificial Intelligence (AI) systems are important components of autonomous vehicles. However, incorporating AI into autonomous vehicles is technically complex, due to the constraints of computation, real-time processing of data, uncertainty handling, and hardware limitations. Bayesian Networks (BNs) are a promising method that allows probabilistic modelling in adaptive learning and environment perception. Here we report on an overview of the application of BNs on autonomous driving, with an emphasis on how BNs can be optimized for embedded system resource constraints, including both computational and energy. Various optimization techniques are discussed, such as model pruning, approximation, acceleration using hardware accelerators, such as Field-Programmable Gate Arrays (FPGA) and Application Specific Integrated Circuit (ASICs), and advanced cooling and power management to ensure AI reliability under high computational load. By reviewing these approaches, we aim to contribute to the development of more robust and green AI systems for autonomous driving.

Luck, skill, and depth of competition in games and social hierarchies.

Patterns of wins and lo sses in pairwise contests, such as occur in sports and games, consumer research and paired comparison studies, and human and animal social hierarchies, are commonly analyzed using probabilistic models that allow one to quantify the strength of competitors or predict the outcome of future contests. Here, we generalize this approach to incorporate two additional features: an element of randomness or luck that leads to upset wins, and a "depth of competition" variable that measures the complexity of a game or hierarchy. Fitting the resulting model, we estimate depth and luck in a range of games, sports, and social situations. In general, we find that social competition tends to be "deep," meaning it has a pronounced hierarchy with many distinct levels, but also that there is often a nonzero chance of an upset victory. Competition in sports and games, by contrast, tends to be shallow, and in most cases, there is little evidence of upset wins.

Refined probabilistic seismic demand modeling of high-speed railway bridge under near-fault hanging-wall effects and velocity pulse effects by a novel ground motion selection approach

Refined probabilistic seismic demand modeling of high-speed railway bridge under near-fault hanging-wall effects and velocity pulse effects by a novel ground motion selection approach

Semantic information guided diffusion posterior sampling for remote sensing image fusion

The task of image fusion for optical images and SAR images is to integrate valuable information from source images. Recently, owing to powerful generation, diffusion models, e.g., diffusion denoising probabilistic model and score-based diffusion model, are flourished in image processing, and there are some effective attempts in image fusion by scholars’ progressive explorations. However, the diffusion models for image fusion suffer from inevitable SAR speckle that seriously shelters from effective information in the same location of optical image. Besides, these methods integrate pixel-level features without information for high-level tasks, e.g., target detection and image classification, which leads fused images are insufficient and their application accuracies are low, for high-level tasks. To tackle these hurdles, we propose the semantic information guided diffusion posterior sampling for image fusion. Firstly, we employ the SAR-BM3D as preprocessing to despeckle. Then, the sampling model is established with fidelity, regularization and semantic information guidance term. The first two terms are obtained by the variational diffusion method via variational inference and first-order stochastic optimization. The last term is served by cross entropy loss between annotation and classification result from FLCNet we design. Finally, the experiments validate the feasibility and superiority of the proposed method on WHU-OPT-SAR dataset and DDHRNet dataset.