Inference Model Research Articles

Statistical inference for lindley random walks with correlated increments

This paper studies statistical inference for a Lindley random walk model when the increment process driving the walk is strictly stationary. Lindley random walks govern customer waiting times in many queueing models and several natural and business processes, including snow depths, frozen soil depths, inventory quantities, etc. The probabilistic properties of a Lindley walk with time-correlated stationary changes are first reviewed. We provide a streamlined argument that the process has a proper limiting distribution when the mean of the incremental changes is negative, and that the Lindley process is strictly stationary when starting from this stationary distribution. Next, the Markov characteristics of the process are explored when the change process has a Markov structure of first or higher order. A derivation of the model's likelihood is given when the change process is a Gaussian autoregressive time series. An efficient particle filtering method for evaluating and optimizing the likelihood with Gaussian changes is then devised and studied via simulation.

Private and Secure Distributed Deep Learning: A Survey

Traditionally, deep learning practitioners would bring data into a central repository for model training and inference. Recent developments in distributed learning, such as federated learning and deep learning as a service (DLaaS) do not require centralized data and instead push computing to where the distributed datasets reside. These decentralized training schemes, however, introduce additional security and privacy challenges. This survey first structures the field of distributed learning into two main paradigms and then provides an overview of the recently published protective measures for each. This work highlights both secure training methods as well as private inference measures. Our analyses show that recent publications while being highly dependent on the problem definition, report progress in terms of security, privacy, and efficiency. Nevertheless, we also identify several current issues within the private and secure distributed deep learning (PSDDL) field that require more research. We discuss these issues and provide a general overview of how they might be resolved.

A hierarchical active inference model of spatial alternation tasks and the hippocampal-prefrontal circuit.

Cognitive problem-solving benefits from cognitive maps aiding navigation and planning. Physical space navigation involves hippocampal (HC) allocentric codes, while abstract task space engages medial prefrontal cortex (mPFC) task-specific codes. Previous studies show that challenging tasks, like spatial alternation, require integrating these two types of maps. The disruption of the HC-mPFC circuit impairs performance. We propose a hierarchical active inference model clarifying how this circuit solves spatial interaction tasks by bridging physical and task-space maps. Simulations demonstrate that the model's dual layers develop effective cognitive maps for physical and task space. The model solves spatial alternation tasks through reciprocal interactions between the two layers. Disrupting its communication impairs decision-making, which is consistent with empirical evidence. Additionally, the model adapts to switching between multiple alternation rules, providing a mechanistic explanation of how the HC-mPFC circuit supports spatial alternation tasks and the effects of disruption.

Parameter estimation for allometric trophic network models: A variational Bayesian inverse problem approach

Abstract Differential equation models are powerful tools for predicting biological systems, capable of projecting far into the future and incorporating data recorded at arbitrary times. However, estimating these models' parameters from observations can be challenging because numerical methods are often required to approximate their solution. An example of such a model is the allometric trophic network model, for which studies considering its inverse problem are limited, particularly in the Bayesian framework. Here we develop a variational Bayesian method for parameter inference of the allometric trophic network model and explore how accurately we can recover its parameter values. We represent the model as a Bayesian neural network, which combines an artificial neural network with Bayesian inference, using a surrogate for the posterior distribution of model parameters, and train this model by evolutionary optimization to avoid potentially costly computation of the gradient with respect to the model parameters. Using synthetic data, we compare the accuracy of this variational inference to ordinary least squares estimation. To reduce the number of estimated parameters, we focus on the inference of functional response parameters. Our variational Bayesian method yields parameter estimates that are comparable to the ordinary least squares results in terms of accuracy. The method provides a promising approach for including uncertainty quantification in parameter estimation, which the simple ordinary least squares approach as it is does not address. Regardless of the method, potential multimodality of the inference problem is nonetheless important to keep in mind. The present study suggests a technique for parameter inference of ordinary differential equation models in the Bayesian context. We propose the method especially for validation of the allometric trophic network model against empirical data.

On the consistent classification and treatment of uncertainties in structural health monitoring applications

Abstract In this paper, we provide a comprehensive definition and classification of various sources of uncertainty within the fields of Structural Dynamics, System Identification, and Structural Health Monitoring (SHM), with a primary focus on the latter. Utilizing the classical input-output system representation as a main contextual framework, we present a taxonomy of uncertainties, intended for consistent classification of uncertainties in SHM applications: (i) input uncertainty; (ii) model form uncertainty; (iii) model parameter/variable uncertainty; (iv) measurement uncertainty; and (v) inherent variability. We then critically review methods and algorithms that address these uncertainties in the context of key SHM tasks: system identification and model inference, model updating, accounting for environmental and operational variability, virtual sensing, damage identification, and prognostic health management. A benchmark shear frame model with hysteretic links is employed as a running example to illustrate the application of selected methods and algorithmic tools. Finally, we discuss open challenges and future research directions in uncertainty quantification for SHM.

Matheamatical Modeling of Multi-Objective Adaptive Neuro Fuzzy Inference Based Optimization for IOT Based Wireless Sensor Network

The proliferation of the Internet of Things (IoT) and its integration with wireless sensor networks (WSNs) necessitates advanced optimization techniques to enhance performance and resource allocation while ensuring reliability and energy efficiency. This study introduces a mathematical modeling approach utilizing a Multi-Objective Adaptive Neuro-Fuzzy Inference System (MO-ANFIS) designed for optimizing IoT-based WSNs. The proposed model synergistically combines the adaptive learning capabilities of neural networks with the reasoning prowess of fuzzy inference systems, encapsulated within a multi-objective framework to concurrently address key operational objectives such as minimizing energy consumption, maximizing network lifetime, and enhancing data accuracy and throughput. The mathematical model is formulated to dynamically adapt to changing network conditions and sensor inputs, enabling real-time tuning of fuzzy rules and membership functions through backpropagation neural training. This adaptability ensures optimal performance despite the variable nature of IoT environments. Simulation results demonstrate that the MO-ANFIS model significantly outperforms traditional optimization methods, offering a robust, scalable solution for complex, dynamic WSNs in the IoT landscape. The findings suggest promising applications in various domains, including smart cities, environmental monitoring, and healthcare, where IoT integration is pivotal. This research not only bridges the gap between theoretical fuzzy-neural frameworks and practical IoT applications but also sets a foundation for future explorations into intelligent, adaptive network management systems.

Assessing the Impact and Cost-Effectiveness of Exposome Interventions on Alzheimer’s Disease: A Review of Agent-Based Modeling and Other Data Science Methods for Causal Inference

Background: The exposome (e.g., totality of environmental exposures) and its role in Alzheimer’s Disease and Alzheimer’s Disease and Related Dementias (AD/ADRD) are increasingly critical areas of study. However, little is known about how interventions on the exposome, including personal behavioral modification or policy-level interventions, may impact AD/ADRD disease burden at the population level in real-world settings and the cost-effectiveness of interventions. Methods: We performed a critical review to discuss the challenges in modeling exposome interventions on population-level AD/ADRD burden and the potential of using agent-based modeling (ABM) and other advanced data science methods for causal inference to achieve this. Results: We describe how ABM can be used for empirical causal inference modeling and provide a virtual laboratory for simulating the impacts of personal and policy-level interventions. These hypothetical experiments can provide insight into the optimal timing, targeting, and duration of interventions, identifying optimal combinations of interventions, and can be augmented with economic analyses to evaluate the cost-effectiveness of interventions. We also discuss other data science methods, including structural equation modeling and Mendelian randomization. Lastly, we discuss challenges in modeling the complex exposome, including high dimensional and sparse data, the need to account for dynamic changes over time and over the life course, and the role of exposome burden scores developed using item response theory models and artificial intelligence to address these challenges. Conclusions: This critical review highlights opportunities and challenges in modeling exposome interventions on population-level AD/ADRD disease burden while considering the cost-effectiveness of different interventions, which can be used to aid data-driven policy decisions.

Bad Local Minima Exist in the Stochastic Block Model

AbstractWe study the disassortative stochastic block model with three communities, a well-studied model of graph partitioning and Bayesian inference for which detailed predictions based on the cavity method exist (Decelle et al. in Phys Rev E 84:066106, 2011). We provide strong evidence that for a part of the phase where efficient algorithms exist that approximately reconstruct the communities, inference based on maximum a posteriori (MAP) fails. In other words, we show that there exist modes of the posterior distribution that have a vanishing agreement with the ground truth. The proof is based on the analysis of a graph colouring algorithm from Achlioptas and Moore (J Comput Syst Sci 67:441–471, 2003).

Shufflenetv2UNet: An improved neural network model for grassland sample coverage extraction

AbstractAccurate extraction of grassland sample coverage is crucial for regional ecological environment monitoring. Due to the strong feature learning capability, high flexibility, and scalability of deep learning methods, they have great potential in grassland sample extraction modelling. However, we still lack a model that can achieve both lightweight structure and effective performance for small object segmentation to considering the small target characteristics of grassland vegetation and the requirements for model deployment in later stages. Here, we combined the UNet model, which performs well in small target segmentation, with the lightweight network Shufflenetv2 model, proposing an improved UNet neural network, Shufflenetv2UNet, for grassland sample coverage extraction. The core of Shufflenetv2UNet is the removal of maximum pooling and double‐layer convolution modules from downsampling in the UNet neural network. In addition, the Inverted Residual Block structure module from Shufflenetv2 was added to achieve a lightweight model and improved extraction accuracy. The Shufflenetv2UNet achieves an accuracy of 98.23%, with a parameter size of 50.74 M, and a model inference speed of 0.004 s. Compared to existing extraction methods, this model has advantages in prediction accuracy, parameter size, and model inference speed. Moreover, Shufflenetv2UNet achieved different types of grassland sample coverage extractions, with good robustness, generalization, and universality, enabling investigators to quickly and accurately obtain grassland sample coverage. This allows more dynamic and accurate ground measurement data for regional grassland environmental monitoring.

Application of Traffic Cone Target Detection Algorithm Based on Improved YOLOv5

To improve the automation level of highway maintenance operations, the lightweight YOLOv5-Lite-s neural network was deployed in embedded devices to assist an automatic traffic cone retractor in completing recognition and positioning operations. The system used the lightweight shuffle Net network as a backbone for feature extraction, replaced convolutional layers with focus modules to reduce computational complexity, and reduced the use of the C3 layer to increase network speed, thereby meeting the speed and accuracy requirements of traffic cone placement and retraction operations while maintaining acceptable model inference accuracy. The experimental results show that the network could maintain recognition accuracy and speed values of around 89% and 9 fps under different working conditions such as varying distances, lighting conditions, and occlusions, meeting the technical requirements for deploying and retrieving cones at a speed of 30 cones per minute when the operating vehicle’s speed was 20 km/h. The automatic traffic cone placement and retraction system operated accurately and stably, achieving the application of machine vision in traffic cone retraction operations.

SpatialGEV: Fast Bayesian inference for spatial extreme value models in R

SpatialGEV: Fast Bayesian inference for spatial extreme value models in R

A fuzzy-logic approach for longitudinal assessment of patients’ psychophysiological state: an application to upper-limb orthopedic robot-aided rehabilitation

Understanding the psychophysiological state during robot-aided rehabilitation is crucial for assessing the patient experience during treatments. This paper introduces a psychophysiological estimation approach using a Fuzzy Logic inference model to assess patients’ perception of robots during upper-limb robot-aided rehabilitation sessions. The patients were asked to perform nine cycles of 3D point-to-point trajectories toward different targets at varying heights with the assistance of an anthropomorphic robotic arm (i.e. KUKA LWR 4+). Physiological parameters, including galvanic skin response, heart rate, and respiration rate, were monitored across ten out of forty daily sessions. This data enabled the construction of an inference model to estimate patients’ perception states. Results expressed in terms of correlation coefficients between the patient state and the increasing number of sessions. Correlation coefficients showed statistically significant strong associations: a state of heightened engagement (formerly described as “Excited”) had ρ=-0.73\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\rho = -0.73$$\\end{document} (p-value=0.01), and a more calm and resting state (namely “Relaxed” state) had ρ=0.70\\documentclass[12pt]{minimal} \\usepackage{amsmath} \\usepackage{wasysym} \\usepackage{amsfonts} \\usepackage{amssymb} \\usepackage{amsbsy} \\usepackage{mathrsfs} \\usepackage{upgreek} \\setlength{\\oddsidemargin}{-69pt} \\begin{document}$$\\rho = 0.70$$\\end{document} (p-value=0.02) with the number of sessions completed. All patients had positive interaction with the robot, initially expressing curiosity and interest that gradually shifted to a more “Relaxed” state over time.

Modified intrinsic Bayes factor for multivariate regression models

The multivariate Regression model (MRM) is one of the most widely used models for multivariate statistical inference. In this paper, a closed-form and fully objective Bayes factor for comparing two MRMs is introduced. We also introduce a Bayes factor-based method for comparing several MRMs. It is well known that the arithmetic intrinsic Bayes factor (AIBF) does not satisfy the pairwise and multiple coherency conditions. So, the AIBF cannot be used in multiple hypotheses testing and model selection. Even in the case of only two competing hypothesis or models, lack of pairwise coherency leads to the embarrassing position of two Bayes factors for comparing two models instead of just one. In this paper, we have made a simple modification to the original definition and introduced the modified intrinsic Bayes factor (ModIBF) which satisfies the pairwise and the multiple coherency conditions.

Transdiagnostic Anxiety-Related Increases in Information Sampling are Associated With Altered Valuation.

Excessive information sampling in psychiatric patients characterized by high trait anxiety has been inconsistently linked with alterations in inferential and valuation processes. Methodological limitations could account in part for these inconsistencies. To address this, computational models of inference and valuation were applied to data collected from a transdiagnostic sample of adults with and without an anxiety or compulsive disorder using a version of the beads task with enhanced experimental controls. Participants diagnosed with an anxiety or compulsive disorder (n = 35) and healthy controls (n = 23) completed the beads task with three majority-to-minority ratios of blue versus green beads (60:40, 75:25, 90:10). First, a Bayesian belief-updating model was fit to quantify the iterative process by which new information (bead color) and prior beliefs were integrated to influence current beliefs about jar identity. Next, a parameterized partially observable Markov decision process model was used to parse the contribution of value-based decisions to sampling behavior and included a relative subjective cost parameter, Csub , for each bead-ratio condition. Higher trait anxiety was associated with more draws-to-decision, most robustly in the 90:10 bead-ratio condition. Only relative subjective cost of sampling decisions, and not inferential differences in weighting of new or old information, satisfactorily accounted for this relation. Specifically, lower Csub(0.9) was associated with more trait anxiety and more draws-to-decision. In a condition with high objective evidence strength, transdiagnostic trait-anxiety-related increases in information sampling were explained by a cost-benefit analysis where relatively higher subjective cost was assigned to an incorrect guess, highlighting valuation as a potential treatment target for future research.

Voice or text? The role of physician media choice on patient experience in online medical communities

AbstractOnline medical communities (OMCs) are a type of online healthcare, in which physician‐patient interaction can be comprised of a variety of media options such as pictures, text, and voice. These media formats are often used to create a personalized patient experience in AI‐driven conversational healthcare platforms. To explore how physician media usage affects patient experience, we propose a counterfactual causal inference model using AI‐driven data mining methods on 131,083 online consultation records and 7,666,111 messages sent by physicians from one of the largest OMCs in China. Our study reveals the negative impact of physician use of voice on patient experience, compared to text. Drawing upon social support theory, we identify the mechanism by which physician media usage for voice produces a negative effect. The findings indicate that the negative effect of physicians' voice‐media usage occurs mainly in low‐risk disease conditions, by weakening the role of professional and emotional support. In contrast, in high‐risk disease conditions, voice‐media usage strengthens the role of professional and emotional support in improving the patient's experience. Our study is one of the first to focus on the social support attributes of the different media formats used in OMCs. We use advanced AI text‐analysis algorithms to extract features related to social support in physician‐patient conversations, and thus contribute to the use of AI in feature extraction for research.

Damage Diagnostics of Miter Gates Using Domain Adaptation and Normalizing Flow-Based Likelihood-Free Inference

Miter gates are vital civil infrastructure components in inland waterway transportation networks. To provide risk-informed insights for decisions related to repair and maintenance, sensors have been installed on some miter gates for monitoring. Despite the monitoring system's ability in collecting a large volume of monitoring data, accurately diagnosing damage state in such large structures remains challenging due to the lack of labeled monitoring data, since these structures are designed with high reliability and for a long operation life. This paper addresses this challenge by proposing a damage diagnostics approach for miter gates based on domain adaptation. The proposed approach consists of two main modules. In the first module, Cycle-Consistent generative adversarial network (CycleGAN) is employed to map monitoring data of a miter gate of interest and other similar yet different miter gates into the same analysis domain. Subsequently, a normalizing flow-based likelihood-free inference model is constructed within this common domain using data from source miter gates whose damage states are labeled from historical inspections. The trained normalizing flow model is then used to predict the damage state of the target miter gate based on the translated monitoring data. A case study is presented to demonstrate the effectiveness of the proposed method. The results indicate that the proposed method in general can accurately estimate the damage state of the target miter gate in the presence of uncertainty.

Power spectrum multipoles and clustering wedges during the Epoch of Reionization

Abstract We study the viability of using power spectrum clustering wedges as summary statistics of 21 cm surveys during the Epoch of Reionization (EoR). For observations in a wide redshift range z ∼ 7 − 9 corresponding to a line-of-sight scale of ∼500 Mpc, the power spectrum is subject to anisotropic effects due to the evolution along the light-of-sight. Information on the physics of reionization can be extracted from the anisotropy using the power spectrum multipoles. Signals of the power spectrum monopole are highly correlated at scales smaller than the typical ionization bubble, which can be disentangled by including higher-order multipoles. By simulating observations of the low frequency part of the Square Kilometre Array (SKA) Observatory, we find that the sampling of the cylindrical wavenumber k-space is highly non-uniform due to the baseline distribution, i.e. the distribution of antenna pairs sampling different transverse k⊥ scales. Measurements in clustering wedges partition the cylindrical k-space into different radial k∥ scales, and can be used for isolating parts of k-space with relatively uniform sampling, allowing for more precise parameter inference. Using Fisher Matrix forecasts, we find that the reionization model can be inferred with per-cent level precision with ∼120 hrs of integration time using SKA-Low. Compared to model inference using only the power spectrum monopole above the foreground wedge, model inference using multipole power spectra in clustering wedges yields a factor of ∼3 improvement on the marginalised 1D parameter constraints.

Probability Modeling and Statistical Inference in Cancer Screening

Probability Modeling and Statistical Inference in Cancer Screening

Active learning strategies for the design of sustainable alloys.

Active learning comprises machine learning-based approaches that integrate surrogate model inference, exploitation and exploration strategies with active experimental feedback into a closed-loop framework. This approach aims at describing and predicting specific material properties, without requiring lengthy, expensive or repetitive experiments. Recently, active learning has shown potential as an approach for the design of sustainable materials, such as scrap-compatible alloys, and for enhancing the longevity of metallic materials. However, in-depth investigations into suited best-practice strategies of active learning for sustainable materials science are still scarce. This study aims to present and discuss active learning strategies for developing and improving sustainable alloys, addressing single-objective and multi-objective learning and modelling scenarios. As model cases, we discuss active learning strategies for optimizing Invar and magnetic alloys, representing single-objective scenarios, and more general steel design approaches, exemplifying multi-objective optimization. We discuss the significance of finding the right balance between exploitation and exploration strategies in active learning and suggest strategies to reduce the number of iterations across diverse scenarios. This kind of research aims to find metrics for a more effective application of active learning and is used here to advance the field of sustainable alloy design.This article is part of the discussion meeting issue 'Sustainable metals: science and systems'.

Real time optimization of distillation columns using data‐driven models

AbstractThis work presents a data‐driven model of a two‐product distillation tower that is suitable for real‐time optimization (RTO) of distillation columns. The proposed model accurately predicts product mass fractions using operating variables and tray temperatures by integrating a linear data‐driven inferential composition model (based on two tray temperatures in each section of the tower, reflux/distillate ratio, and reboiler duty/bottoms flow ratio) with a neural network (NN) model that predicts tray temperatures from the value of the manipulated variables. RTO is carried out via an iterative procedure where the sensitivity matrix is initially calculated from the model and updated using plant measurements from subsequent values. A butane splitter column is presented as a case study. Our results show that the implementation of the data‐driven model‐based RTO results in a solution that is within 0.1% of the optimization solution based on the rigorous tray‐to‐tray distillation simulation.