Linear Gaussian Research Articles

Machine learning-based optimization of photogrammetric JRC accuracy

To improve the accuracy of photogrammetric joint roughness coefficient (JRC) estimation, this study proposes two optimization models based on ground sample distance (GSD), point density, and the root mean square error (RMSE) of checkpoints. First, an algorithm that automatically generates spatial positions for equipment based on the convergence strategy was developed, using principles of Structure from Motion and Multi-View Stereo (SfM-MVS) and the shooting parameter selection algorithm (SPSA). Second, a portable positioning plate containing ground control points and checkpoints was designed based on optical principles, and a moving camera capture strategy guided by SPSA was proposed. Combining SPSA, portable positioning plate, and moving camera capture strategy, a photogrammetric experiment for small-scale rock samples in the field was conducted, collecting 48 datasets with different shooting parameters. Subsequently, a dataset incorporating GSD, point density, RMSE, and three JRC estimation metrics was established, revealing their correlations and sensitivities. Using seven machine learning algorithms, optimization models for photogrammetric JRC accuracy were developed, with Linear Multidimensional Regression and Gaussian Process Regression models improving JRC accuracy by an average of 85.73%. Finally, the applicability and limitations of the newly proposed method were further discussed.

Moving Target Detection in Clutter Environment Based on Track Posture Hypothesis Testing

ABSTRACTMoving target detection in a heavy clutter area is a challenging problem in the field of radar automatic target detection. Within the framework of track posture hypothesis testing, the radar‐blips‐oriented target detection algorithms named ErgSad and RanSaD are proposed for single target and formation targets detection, respectively in this paper. With the assumption that target motion is a linear Gaussian process in a short time, the algorithms use dual blips in different scans to generate the track hypothesis and test the hypothesis with residual blips in other scans to determine the existence of moving targets. To reduce time consumption and minimize the error of model estimation, the sampling rules and the supporting domain related to timestamps of track hypothesis models were designed. Simulation data experiments show that the proposed algorithms have more superior performance to detect targets than the state‐of‐the‐art algorithms in the serious clutter environment.

Partially observable optimal control using exponential cost criterion

We consider a partially observable optimal control model in which the system and observation noises are correlated, and the cost criterion is in the form of an exponential of an integral. A new minimum principle criterion is derived and applied to compute the explicit solution of linear exponential Gaussian optimisation problem with correlated noises.

Optimal quantized feedback control for linear quandratic Gaussian systems with input delay

AbstractThis paper is concerned with the optimal quantized feedback linear quadratic Gaussian (LQG) control problem for a discrete‐time stochastic system with input delay as well as the measurements to be quantized before transmitted to the controller. In this scenario, the system is presented with several choices of quantizers, along with the cost of using each quantizer. The objective is to jointly select the quantizers and synthesize the controller to maintain an optimal balance between control performance and quantization cost. It is shown that this problem can be decoupled into two optimization problems when the innovation signal is quantized instead of state: one for optimal controller synthesis and the other for optimal quantizer selection. More specifically, a necessary and sufficient condition is derived for the optimal control problem based on Pontryagin's maximum principle. On the other hand, the optimal quantizer selection policy is established by dealing with a certain Markov decision process (MDP).

TAMie Force Field for Alkanethiols: Multifidelity Gaussian Processes for Dealing with Scarce Experimental Data.

This study extends the transferable anisotropic Mie potential (TAMie) to alkanethiols. The force field parameters are optimized by using an analytic equation of state as a surrogate model. Given the lack of experimental density data at elevated temperatures where Monte Carlo simulations have high statistical precision, the equation of state is supplemented by a linear multifidelity Gaussian process approach to bridge the temperature gap. Force field parameters are adjusted by minimizing squared deviations of calculated vapor pressures and liquid densities from experimental data of 1-propanethiol, 1-butanethiol and 1-pentanethiol leading to small mean absolute relative deviations in liquid densities and vapor pressures. The force field is transferable to higher 1-thiols, as shown for 1-hexanethiol and 1-octanethiol. Individual parameter sets are provided for methanethiol and ethanethiol. The shear viscosity of pure substances is predicted in fair agreement with experimental data, considering that it is not included in the parametrization. Further, the phase behavior of binary mixtures of alkanethiols with alkanes is studied, and predictions of the TAMie model are found in excellent agreement with experimental data.

On a class of linear quadratic Gaussian quantilized mean field games

An energy provider faced with energy generation risks and a large homogeneous pool of customers designs its energy price as a time-varying function of a risk-related quantile of the total energy demand, which generalizes pricing through the mean of the total energy demand. In the infinite population limit, we model the pricing problem with a class of linear quadratic Gaussian quantilized mean field games. For these quantilized mean field games, we show existence and uniqueness of an equilibrium which reveals the price trajectory, as well as an approximate Nash property when the quantilized mean field game’s feedback control functions are applied to the large but finite game and the rate of convergence of the Nash deviation to zero as a function of the population size and the quantile is provided. Finally, the use of this class of quantilized mean field games is illustrated in the context of equivalent thermal parameter models for households heater and an energy provider using solar generation.

Adversarial Missingness Attacks on Causal Structure Learning

Causality-informed machine learning has been proposed as an avenue for achieving many of the goals of modern machine learning, from ensuring generalization under domain shifts to attaining fairness, robustness, and interpretability. A key component of causal machine learning is the inference of causal structures from observational data; in practice, this data may be incompletely observed. Prior work has demonstrated that adversarial perturbations of completely observed training data may be used to force the learning of inaccurate causal structural models (SCMs). However, when the data can be audited for correctness (e.g., it is cryptographically signed by its source), this adversarial mechanism is invalidated. This work introduces a novel attack methodology wherein the adversary deceptively omits a portion of the true training data to bias the learned causal structures in a desired manner (under strong signed sample input validation, this behavior seems to be the only strategy available to the adversary). Under this model, theoretically sound attack mechanisms are derived for the case of arbitrary SCMs, and a sample-efficient learning-based heuristic is given. Experimental validation of these approaches on real and synthetic data sets, across a range of SCMs from the family of additive noise models (linear Gaussian, linear non-Gaussian, and non-linear Gaussian) demonstrates the effectiveness of adversarial missingness attacks at deceiving popular causal structure learning algorithms.

Erosive wear and particle attrition in multi-stage solar particle receivers and screw conveyors: A CFD-DEM approach with machine learning and artificial neural networks

A coupled finite volume and discrete element method (CFD-DEM) numerical model is developed to unravel the fundamental mechanisms of granular transport, surface erosive wear and particle attrition in multistage solar particle receivers (MSPR) and screw conveyors (SC). Additionally, various regression-based supervised machine learning (ML) and artificial neural networks (ANN) are trained and optimized with the goal of quickly quantifying erosion and attrition thereby overcoming (a) the inherent challenges of high computational expense of CFD-DEM simulations and (b) exorbitant time required to post-process and extract DEM results. This study is the first of its kind to harness a synergistic numerical framework (CFD-DEM, ML, ANN) to unravel the fundamental mechanisms of erosion and attrition in MSPRs, non-vibrationally induced SCs, and vibrationally-induced SCs. It is found that the packing distribution, particle curtain width, and severity of particle scattering in a MSPR exhibits a subtle variation with particle diameter. In the context of SCs, the magnitudes of erosion and attrition are contingent on the particle diameter and flow operating conditions such as screw blade velocity, vibrational frequency and vibrational amplitude. Interestingly, a vibrationally-induced SC (VI-SC) exhibits superior performance in terms of low erosion and attrition unlike non-vibrationally induced SC (NVI-SC). It is found that MSPRs with 750–1000 µm particles and SCs with 750–1000 µm coupled with vibrational frequencies and vibrational amplitudes of 5 Hz and 0.007 m, respectively, are the preferred configurations due to minimum particle scattering, low erosive wear and low particle attrition. The trained and optimized ML models such as XGBoost, CatBoost, and Multilayer Perceptron (MLP) exhibit superior performance in predicting erosive wear and particle attrition on unseen DEM data, whereas the Multiple Linear Regression (MLR) and Gaussian Process Regression (GPR) ML exhibit poor performance prediction. For the first time, new information on fundamental granular flow dynamics, erosive wear and particle attrition in MSPRs and SCs is discerned. The methodology allows scientists to quickly examine material degradation for a myriad of industrial applications such as concentrated solar thermal (CST), among others.

Microwave-based and convective drying of cabbage (Brassica oleracea L. var capitata L.): Computational intelligence modeling, thermophysical properties, quality and mid-infrared spectrometry

This study assessed and compared the impact of hot air oven drying (HAD) at 50, 60 and 70 °C and microwave drying (MWD) at 195, 307 and 521 W on the kinetics of thermophysical properties and quality of dried cabbage. The thermophysical properties were computed using established model equations. The proximate composition, bioactive compounds, antioxidant capacity (AOC), color and functional groups were analyzed using Association of official analytical chemists (AOAC) test protocol, High performance liquid chromatography (HPLC), Ultraviolet-visible (UV–Vis) spectrometry and Attenuated total reflectance coupled to Fourier transformed infrared spectroscopy (ATR-FTIR), respectively. The moisture ratio was modeled using probabilistic computational intelligence modeling approaches. The results revealed that only density, thermal diffusivity and thermal effusivity appeared to be kinetically controlled by moisture content. Stepwise linear regression (SLR) model and Gaussian process regression (GPR) model were more accurate in simulating the moisture ratio. Increasing HAD temperature preserved all of the proximate compositions except carbohydrates. Similarly, increasing MWD power preserved all the proximate compositions except crude fiber. Increasing HAD temperature from 50 to 60 °C and 60–70 °C decreased the total flavonoid content (TFC) and total phenolic content (TPC) by 21.31 % and 13.60 %, respectively but increased the AOC from 27.2 to 39.6 %. Increasing MWD power levels from 195 to 307 W and 307–521 W decreased TFC by 13.2 % and 0.7 %, respectively but increased the TPC by 9.72 % and 13.14 %, thus AOC increased from 35.2 to 45.7 %. MWD presented higher AOC relative to HAD. Drying at 50 °C and 307 W gave the highest whitening index in HAD and MWD. ATR-FTIR analysis revealed the formation of new compounds by thermal effect. SLR and GPR were more accurate in modeling the drying kinetics. HAD and MWD affected the quality parameters of cabbage differently, thus the choice of the dryer type will depend on the parameters of interest.

Additive dynamic Bayesian networks for enhanced feature learning in soft sensor modeling

Due to the advantages of indicating variable structure and efficient reasoning, Bayesian Networks (BN) have been widely used in data-driven soft sensor applications. However, restricted to linear and conditional Gaussian property, BN-based soft sensors rarely achieve high prediction accuracy. In this paper, an ensemble learning framework – additive dynamic Bayesian networks (ADBN) is proposed for enhanced feature learning, in which Dynamic Bayesian networks are used to learn the conditional independencies among variables and construct feature sets for the following base learners. Additional DBNs are constructed upon the residual information from the past model, to carry out feature learning to fit the residuals. The procedure is repeated and a termination rule from the perspective of feature learning is proposed to end this process, and thus the model complexity can be well restricted. The proposed method is validated on two actual industrial cases. It reveals that the ADBN feature learning method has obtained great improvements. Compared to the single DBN feature engineering method, the root mean square error (RMSE) performance has been improved by 20% and 13% on the two cases, respectively.

Propensity Weighted federated learning for treatment effect estimation in distributed imbalanced environments

Estimating treatment effects from observational data in medicine using causal inference is a very relevant task due to the abundance of observational data and the ethical and cost implications of conducting randomized experiments or experimental interventions. However, how could we estimate the effect of a treatment in a hospital that has very restricted access to treatment? In this paper, we want to address the problem of distributed causal inference, where hospitals not only have different distributions of patients, but also different treatment assignment criteria. Furthermore, it is necessary to take into account that due to privacy restrictions, personal patient data cannot be shared between hospitals. To address this problem, we propose an adaptation of the federated learning algorithm FederatedAveraging to one of the most advanced models for the prediction of treatment effects based on neural networks, TEDVAE. Our algorithm adaptation takes into account the shift in the treatment distribution between hospitals and is therefore called Propensity WeightedFederatedAveraging (PW FedAvg). As the distributions of the assignment of treatments become more unbalanced between the nodes, the estimation of causal effects becomes more challenging. The experiments show that PW FedAvg manages to reduce errors in the estimation of individual causal effects when imbalances are large, compared to VanillaFedAvg and other federated learning-based causal inference algorithms based on the application of federated learning to linear parametric models, Gaussian Processes and Random Fourier Features.

Linear Gaussian bounding box representation and ring-shaped rotated convolution for oriented object detection

In oriented object detection, current representations of oriented bounding boxes (OBBs) often suffer from the boundary discontinuity problem. Methods of designing continuous regression losses do not essentially solve this problem. Although Gaussian bounding box (GBB) representation avoids this problem, directly regressing GBB is susceptible to numerical instability. We propose linear GBB (LGBB), a novel OBB representation. By linearly transforming the elements of GBB, LGBB avoids the boundary discontinuity problem and has high numerical stability. In addition, existing convolution-based rotation-sensitive feature extraction methods only have local receptive fields, resulting in slow feature aggregation. We propose ring-shaped rotated convolution (RRC), which adaptively rotates feature maps to arbitrary orientations to extract rotation-sensitive features under a ring-shaped receptive field, rapidly aggregating features and contextual information. Experimental results demonstrate that LGBB and RRC achieve state-of-the-art performance. Furthermore, integrating LGBB and RRC into various models effectively improves detection accuracy.

Analysis and Study of Transmission Line Icing Based on Grey Correlation Pearson Combinatorial Optimization Support Vector Machine

In recent years, ice and snow disasters have occurred frequently, and transmission line icing can pose a great threat to the operation of the power system. In order to solve this problem, the power industry should introduce an intelligent monitoring system to recognize the risk of ice-covered disasters in advance and take timely countermeasures to reduce the impact of extreme weather on the safe operation of the power system. In this paper, we first use Python to sort out and summarize multidimensional meteorological data from multiple locations in the Northeast China, and propose a method called Grey Relational Pearson Combination Analysis (GPCA). Through this method, we determine the correlation between multidimensional meteorological data and icing degree, and select the optimal combination of icing factors as the feature vector for a support vector machine. In addition, we also compare the differences in classification effects of four types of kernel functions, such as linear kernel, polynomial kernel and Gaussian kernel. Finally, considering data balance, we use Gaussian kernel to establish GPCA-SVM model and conduct experimental analysis. There was a 1.04% increase in accuracy, a 1.36% increase in recall, and 0.31% increase in F1 value when comparing gray correlation accuracy and a 21.66% increase in precision when comparing Pearson accuracy at light icing. Comparison of gray correlation accuracy at heavy icing improved by 1.04%, precision by 17.44%, and F1 value by 10.29%; comparison of Pearson accuracy improved by 0.31%, precision by 9.93%, and F1 value by 1.67%. These research results lay the foundation for the intelligent monitoring system of power grid, and provide guarantee for the operation and maintenance of power system.

Immunotherapy toxicity prediction in patients with melanoma using machine learning algorithms.

e21567 Background: Immune checkpoint inhibitor (ICI) related toxicity is common in melanoma patients, but identifying who will experience toxicity can be challenging. Precision medicine tools that accurately predict ICI toxicity could improve patient outcomes. This study aimed to use Machine Learning (ML) to predict ICI toxicity in patients with melanoma. Methods: 384 datasets were available for patients started on immunotherapy between 2014-2023 in a single, large regional cancer center in Kent, U.K. Raw data was preprocessed using a standard scaling function and one hot encoding, followed by SMOTE to balance the imbalanced classes. Six ML algorithms were developed using 80% of the training data, tested on 20% of validation data and used the Grid Search Cross-Validation technique for hyperparameter optimization. Shapley Additive Explanations (SHAP) explainable artificial intelligence was used to interpret the toxicity prediction model to improve the interpretation and transparency of the decision boundary. Results: Of the 384 patients, 112 patients (29%) had completely resected disease and then had adjuvant treatment while 272 (71%) were unresectable or had metastatic disease and had palliative treatment intent. In the metastatic setting, the commonest immunotherapy regimen was Pembrolizumab (155, 57%), followed by Ipilimumab/Nivolumab (85, 31%). The commonest regimen in the adjuvant setting was Pembrolizumab (86, 77%). 95 patients (25%) experienced ICI-related toxicity. The number of patients with Grades 1-4 toxicity was 2 ,4, 18 and 4 respectively, the remaining had unspecified toxicity grades. The commonest class of toxicity (CTCAE v5) was gastrointestinal (36, 38%), followed by endocrine (20, 21%) and skin (19, 20%). Linear Regression and Gaussian Naïve Bayes predicted toxicity with the most accuracy, 92% (Table 1). The model showed that increasing age, female gender and not receiving Nivolumab predicted toxicity. Regardless of the regimen or treatment intent, our ML model could also predict immunotherapy response with 90% accuracy. Conclusions: These ML algorithms used clinically available data to predict ICI toxicity accurately. Key predictors of toxicity included increasing age, female gender and not receiving Nivolumab in the setting of both adjuvant and palliative intent. Future work will aim to externally validate these models in other centers. We will also explore if models could be trained for use in patients treated with nivolumab in combination with new anti-LAG-3(relatlimab). [Table: see text]

On learning time series DAGs: A frequency domain approach

The fields of time series and graphical models emerged and advanced separately. Previous work on the structure learning of continuous and real-valued time series utilizes the time domain, with a focus on either structural autoregressive models or linear (non-)Gaussian Bayesian Networks. In contrast, a novel frequency domain approach is proposed to identify a topological ordering and learn the structure of multivariate time series. In particular, a class of complex-valued Structural Causal Models (cSCM) is defined at each frequency of the Fourier transform of the time series. Assuming that the time series is generated from the transfer function model, it is demonstrated that the topological ordering and the corresponding summary directed acyclic graph can be uniquely identified from cSCM. The performance of the algorithm is investigated using simulation experiments and real datasets.

A lower bound estimate of life span of solutions to stochastic 3D Navier–Stokes equations with convolution-type noise

In this paper, we investigate the stochastic 3D Navier–Stokes equations perturbed by linear multiplicative Gaussian noise of convolution type by transformation to random PDEs. We focus on obtaining bounds from below for the life span associated with regular initial data. The key point of the proof is the fixed point argument.

Kalman filter with impulse noised outliers: a robust sequential algorithm to filter data with a large number of outliers.

Impulse noised outliers are data points that differ significantly from other observations. They are generally removed from the data set through local regression or the Kalman filter algorithm. However, these methods, or their generalizations, are not well suited when the number of outliers is of the same order as the number of low-noise data (often called nominal measurement). In this article, we propose a new model for impulsed noise outliers. It is based on a hierarchical model and a simple linear Gaussian process as with the Kalman Filter. We present a fast forward-backward algorithm to filter and smooth sequential data and which also detects these outliers. We compare the robustness and efficiency of this algorithm with classical methods. Finally, we apply this method on a real data set from a Walk Over Weighing system admitting around 60 % of outliers. For this application, we further develop an (explicit) EM algorithm to calibrate some algorithm parameters.

Linear Deconfounded Score Method: Scoring DAGs With Dense Unobserved Confounding.

This article deals with the discovery of causal relations from a combination of observational data and qualitative assumptions about the nature of causality in the presence of unmeasured confounding. We focus on applications where unobserved variables are known to have a widespread effect on many of the observed ones, which makes the problem particularly difficult for constraint-based methods, because most pairs of variables are conditionally dependent given any other subset, rendering the causal effect unidentifiable. In this article, we show that under the principle of independent mechanisms, unobserved confounding in this setting leaves a statistical footprint in the observed data distribution that allows for disentangling spurious and causal effects. Using this insight, we demonstrate that a sparse linear Gaussian directed acyclic graph (DAG) among observed variables may be recovered approximately and propose a simple adjusted score-based causal discovery algorithm that may be implemented with general-purpose solvers and scales to high-dimensional problems. We find, in addition, that despite the conditions we pose to guarantee causal recovery, performance in practice is robust to large deviations in model assumptions, and extensions to nonlinear structural models are possible.

Data-driven online tracking filter architecture: A LightGBM implementation

State estimation has been dominated by traditional algorithms based on the Bayesian framework, and Bayesian filtering has proven to be an effective component with a wide range of applications in various fields. On the other hand, Bayesian framework is inherently limited by the precise model prior information, while struggling to cope with the challenges such as Markov assumptions that do not adequately describe the actual motion modes, and model transition decision delay in the multiple model algorithm. To alleviate this, we redefine the state filtering problem from a data-driven perspective as a non-probabilistic mapping problem from measurement sequence to target state, replacing the iteration of model derivation with supervised learning regression. Specifically, we introduce a data-driven tracking filter architecture (DTF) that includes spatio-temporal feature processing and decoupling of target trend and sensor uncertainty features. Its fast implementation based on LightGBM is further given, considering the interpretability of the model. Simulation results show that the proposed method approaches to the theoretical optimal solution of the Kalman filter in an ideal linear Gaussian scenario. In challenging maneuvering target tracking scenarios, the proposed model is superior in performance compared with other existing methods. Finally, we validate the effectiveness of the method in radar tracking and real-world video scenes.

An efficient breast cancer classification model using bilateral filtering and fuzzy convolutional neural network

BC (Breast cancer) is the second most common reason for women to die from cancer. Recent workintroduced a model for BC classifications where input breast images were pre-processed using median filters for reducing noises. Weighed KMC (K-Means clustering) is used to segment the ROI (Region of Interest) after the input image has been cleaned of noise. Block-based CDF (Centre Distance Function) and CDTM (Diagonal Texture Matrix)-based texture and shape descriptors are utilized for feature extraction. The collected features are reduced in counts using KPCA (Kernel Principal Component Analysis). The appropriate feature selection is computed using ICSO (Improved Cuckoo Search Optimization). The MRNN ((Modified Recurrent Neural Network)) values are then improved through optimization before being utilized to divide British Columbia into benign and malignant types. However, ICSO has many disadvantages, such as slow search speed and low convergence accuracy and training an MRNN is a completely tough task. To avoid those problems in this work preprocessing is done by bilateral filtering to remove the noise from the input image. Bilateral filter using linear Gaussian for smoothing. Contrast stretching is applied to improve the image quality. ROI segmentation is calculated based on MFCM (modified fuzzy C means) clustering. CDTM-based, CDF-based color histogram and shape description methods are applied for feature extraction. It summarizes two important pieces of information about an object such as the colors present in the image, and the relative proportion of each color in the given image. After the features are extracted, KPCA is used to reduce the size. Feature selection was performed using MCSO (Mutational Chicken Flock Optimization). Finally, BC detection and classification were performed using FCNN (Fuzzy Convolutional Neural Network) and its parameters were optimized using MCSO. The proposed model is evaluated for accuracy, recall, f-measure and accuracy. This work’s experimental results achieve high values of accuracy when compared to other existing models.