Parameter Update Research Articles

Bayesian updating of tall timber building model using modal data

A framework for the probabilistic finite element model updating based on measured modal data is presented. The described framework is applied to a seven-storey building made of cross-laminated timber panels. The experimental estimates based on the forced vibration test are used in the process of model updating. First, a generalized Polynomial Chaos surrogate model is derived representing the map from the model parameters to the eigenfrequencies and the eigenvectors. To overcome the difficulties caused by mode switching, we propose a novel approach to mode tracking based on partitioning an extended and low-rank representation of the k mode shapes resulting from different setups of the finite element model into k clusters by the k-means clustering algorithm. Second, the surrogate model derived with the help of mode pairing is used to efficiently perform sensitivity analysis and uncertainty quantification of the first five frequencies and the corresponding mode shapes. Finally, the surrogate-based Bayesian update of the model parameters is efficiently performed, providing engineers not only with a finite element model that gives a good fit to the experimental modal data, but also a stochastic model that represents the uncertainties originating from the initial model and the uncertainties of measuring modal properties.

Impact of DNA Geometry and Scoring on Monte Carlo Track-Structure Simulations of Initial Radiation-Induced Damage.

Track structure Monte Carlo simulations are a useful tool to investigate the damage induced to DNA by ionizing radiation. These simulations usually rely on simplified geometrical representations of the DNA subcomponents. DNA damage is determined by the physical and physicochemical processes occurring within these volumes. In particular, damage to the DNA backbone is generally assumed to result in strand breaks. DNA damage can be categorized as direct (ionization of an atom part of the DNA molecule) or indirect (damage from reactive chemical species following water radiolysis). We also consider quasi-direct effects, i.e., damage originated by charge transfers after ionization of the hydration shell surrounding the DNA. DNA geometries are needed to account for the damage induced by ionizing radiation, and different geometry models can be used for speed or accuracy reasons. In this work, we use the Monte Carlo track structure tool TOPAS-nBio, built on top of Geant4-DNA, for simulation at the nanometer scale to evaluate differences among three DNA geometrical models in an entire cell nucleus, including a sphere/spheroid model specifically designed for this work. In addition to strand breaks, we explicitly consider the direct, quasi-direct, and indirect damage induced to DNA base moieties. We use results from the literature to determine the best values for the relevant parameters. For example, the proportion of hydroxyl radical reactions between base moieties was 80%, and between backbone, moieties was 20%, the proportion of radical attacks leading to a strand break was 11%, and the expected ratio of base damages and strand breaks was 2.5-3. Our results show that failure to update parameters for new geometric models can lead to significant differences in predicted damage yields.

Improving Bayesian Neural Networks by Adversarial Sampling

Bayesian neural networks (BNNs) have drawn extensive interest due to the unique probabilistic representation framework. However, Bayesian neural networks have limited publicized deployments because of the relatively poor model performance in real-world applications. In this paper, we argue that the randomness of sampling in Bayesian neural networks causes errors in the updating of model parameters during training and some sampled models with poor performance in testing. To solve this, we propose to train Bayesian neural networks with Adversarial Distribution as a theoretical solution. To avoid the difficulty of calculating Adversarial Distribution analytically, we further present the Adversarial Sampling method as an approximation in practice. We conduct extensive experiments with multiple network structures on different datasets, e.g., CIFAR-10 and CIFAR-100. Experimental results validate the correctness of the theoretical analysis and the effectiveness of the Adversarial Sampling on improving model performance. Additionally, models trained with Adversarial Sampling still keep their ability to model uncertainties and perform better when predictions are retained according to the uncertainties, which further verifies the generality of the Adversarial Sampling approach.

Updated parameterization of internal tidal mixing in the deep ocean based on laboratory rotating tank experiments

The Coriolis effect plays an important role in the generation and dissipation of internal tides. In this study, the effects of the Coriolis parameter on internal tides generated by tidal current flowing over rough subsurface topography were investigated in the laboratory. The results show that the variation in the internal tidal local dissipation rate with variation in the Coriolis parameter in the deep layer follows a Gaussian distribution. The strongest dissipation occurs near the critical latitude where the local Coriolis parameter is half of the tidal frequency. A modified internal tidal mixing parameterization was developed by introducing a correction function related to the Coriolis effect on internal tidal local dissipation in the deep ocean, based on the original version developed by St. Laurent et al. (2002). The new parameterization was then implemented in a parallel ocean model. The results indicate that the deep global meridional overturning circulation (GMOC) is strengthened in the low and middle latitudes and the Southern Ocean by the change, compared with the original parameterization.

A BERT-Based Aspect-Level Sentiment Analysis Algorithm for Cross-Domain Text.

Cross-domain text sentiment analysis is a text sentiment classification task that uses the existing source domain annotation data to assist the target domain, which can not only reduce the workload of new domain data annotation, but also significantly improve the utilization of source domain annotation resources. In order to effectively achieve the performance of cross-domain text sentiment classification, this paper proposes a BERT-based aspect-level sentiment analysis algorithm for cross-domain text to achieve fine-grained sentiment analysis of cross-domain text. First, the algorithm uses the BERT structure to extract sentence-level and aspect-level representation vectors, extracts local features through an improved convolutional neural network, and combines aspect-level corpus and sentence-level corpus to form a sequence sentence pair. Then, the algorithm uses domain adversarial neural network to make the feature representation extracted from different domains as indistinguishable as possible, that is, the features extracted from the source domain and the target domain have more similarity. Finally, by training the sentiment classifier on the source domain dataset with sentiment labels, it is expected that the classifier can achieve a good sentiment classification effect in both source and target domain, and achieve sentence-level and aspect-level sentiment classification. At the same time, the error pooled values of the sentiment classifier and the domain adversary are passed backwards to realize the update and optimization of the model parameters, thereby training a model with cross-domain analysis capability. Experiments are carried out on the Amazon product review dataset, and accuracy and F1 value are used as evaluation indicators. Compared with other classical algorithms, the experimental results show that the proposed algorithm has better performance.

Adaptive state-feedback stabilisation of output-constrained high-order nonlinear systems with iISS inverse dynamics

This paper investigates the adaptive state-feedback stabilisation for a class of output-constrained high-order nonlinear systems with integral input-to-state stability (iISS) inverse dynamics. The presence of weaker iISS inverse dynamics and parametric uncertainty makes the output-constrained systems in question essentially different from those in the related works. By integrating the tan-type barrier Lyapunov function, sign function and integral Lyapunov functions into the adding a power integrator technique, and choosing the appropriate designed functions and parameter update laws, a unified adaptive state-feedback controller, which can deal with both constrained and unconstrained systems, is constructed. It is proved that all the closed-loop signals are bounded, the symmetric output constraint is not violated, and system states converge to zero asymptotically. Finally, a simulation example is given to illustrate the effectiveness of this control scheme.

A provably efficient monotonic-decreasing algorithm for shape optimization in Stokes flows by phase-field approaches

In this work, we study shape optimization problems in the Stokes flows. By phase-field approaches, the resulted total objective function consists of the dissipation energy of the fluids and the Ginzburg–Landau energy functional as a regularizing term for the generated diffusive interface, together with a Lagrangian multiplier for volume constraint. An efficient decoupled scheme is proposed to implement by the gradient flow approach to decrease the objective function. In each loop, we first update the velocity field by solving the Stokes equation with the phase field variable given in the previous iteration, which is followed by updating the phase field variable by solving an Allen–Cahn-type equation using a stabilized scheme. We then take the cut-off post-processing for the phase-field variable to constrain its value in [0,1]. In the last step of each loop, the Lagrangian parameter is updated with an appropriate artificial time step. We rigorously prove that the proposed scheme permits an unconditionally monotonic-decreasing property. To enhance the overall efficiency of the algorithm, in each loop we update the phase field variable and Lagrangian parameter several time steps but update the velocity field only one time. Numerical results for various shape optimizations are presented to validate the effectiveness of our numerical scheme.

Application analysis of digital fund prediction model based on neural network

SummaryWith the development of the economy, digital funds have developed rapidly, and traditional big data processing technologies have been unable to effectively solve the update of digital funds information. This research uses neural network to study the digital fund forecasting model based on big data technology and deep learning algorithm. First, on the basis of the feedforward neural network model, the network structure and network parameters are optimized to effectively improve the algorithm's ability to learn the characteristics of digital capital information. Through the first‐order approximation, the number of iterations in the model parameter update is effectively reduced, and the computational efficiency of the algorithm is improved. By improving the model network structure, increasing the number of neurons in the hidden layer, and improving the dynamic change learning ability of the algorithm, the dynamic characteristics of the incremental file information can be effectively obtained. Through the experimental simulation calculation, it can be seen that compared with the traditional algorithm, the algorithm proposed in this study has certain advantages in terms of convergence speed, convergence ability and solution error, and the algorithm has good stability and strong inheritance, and is suitable for digital capital forecasting aspects of research.

Peanut anaphylaxis in 2022: Decoupling epinephrine usage from emergency department evaluation

Peanut anaphylaxis in 2022: Decoupling epinephrine usage from emergency department evaluation

Robust correlation filter tracking based on response map analysis network

Conventional discriminative-correlation-filter-based (DCF-based) visual tracking methods always update model at a fixed frequency and learning rate. Without evaluating the tracking confidence scores, the response map generated by filter is the only evidence for locating. Thus, most of the existing DCF-based methods suffer from the model contamination caused by drastic appearance variations, which leads to tracking drift even failure. And excessively frequent update will increase the computational redundancy and risk of over-fitting. In addition, these methods cannot recover target from heavy occlusion neither. Based on the observation that the shape of response maps reflects the matching degree between filter and target, we design and train a small-scale binary network named as response map analysis network (RAN) to evaluate the confidence scores of filters. Further, we propose to learn multiple filters to exploit different kinds of features, and adaptively adjust the update parameters according to the corresponding confidence scores. Moreover, we build a simple occlusion event model to detect heavy occlusion and recover target. Extensive experimental results validate the effectiveness of RAN and demonstrate that the proposed tracker performs favorably against other state-of-the-art (SOTA) DCF-based trackers in terms of precision, overlap rate and efficiency.

Two‐stage degradation modeling for remaining useful life prediction based on the Wiener process with measurement errors

AbstractRemaining useful life prediction (RUL) is a key component in the application of prognostics and health management associated with devices or systems. But such RUL predictions are cumbersome owing to complexities from external effects and internal degradation mechanisms within systems. Specifically, it is common for degradation processes to comprise distinct multiple stages rather than just one uniform stage in many mechanical systems. In particular, the two‐stage degradation modeling for RUL prediction based on the Wiener process with linear drift has received significant attention in recent years. However, negative effects of measurement errors and stochasticity of the degradation states are generally excluded from current degradation modeling, which causes inaccuracy problems that can impact system maintenance schedules and operational efficiency. Therefore, to solve such problems, measurement errors are considered in this paper and a two‐stage degradation model is proposed, in which an adaptive term is also characterized by the Wiener process. The transition probability density function (TPDF) of the degradation state at the two‐stage changing point is derived and an analytical solution for the RUL is obtained under the concept of the first hit time (FHT). A Kalman filter and smoothing algorithm are introduced to estimate variables, and the expectation maximization (EM) algorithm is applied to update and estimate model parameters. Finally, the effectiveness and applicability of the proposed model in RUL predictions are verified through numerical simulation and a case study of bearings.

An Unsupervised Transfer Learning Framework for Visible-Thermal Pedestrian Detection.

Dual cameras with visible-thermal multispectral pairs provide both visual and thermal appearance, thereby enabling detecting pedestrians around the clock in various conditions and applications, including autonomous driving and intelligent transportation systems. However, due to the greatly varying real-world scenarios, the performance of a detector trained on a source dataset might change dramatically when evaluated on another dataset. A large amount of training data is often necessary to guarantee the detection performance in a new scenario. Typically, human annotators need to conduct the data labeling work, which is time-consuming, labor-intensive and unscalable. To overcome the problem, we propose a novel unsupervised transfer learning framework for multispectral pedestrian detection, which adapts a multispectral pedestrian detector to the target domain based on pseudo training labels. In particular, auxiliary detectors are utilized and different label fusion strategies are introduced according to the estimated environmental illumination level. Intermediate domain images are generated by translating the source images to mimic the target ones, acting as a better starting point for the parameter update of the pedestrian detector. The experimental results on the KAIST and FLIR ADAS datasets demonstrate that the proposed method achieves new state-of-the-art performance without any manual training annotations on the target data.

Real-time Parameter Updating and Path-following Control for Autonomous Vehicles on Slope Roads

Real-time Parameter Updating and Path-following Control for Autonomous Vehicles on Slope Roads

On the Development and Performance Evaluation of Improved Radial Basis Function Neural Networks

This article deals with the development of four modified radial basis function neural network (RBFNN) models. The corresponding learning algorithms associated with the updating of internal parameters of the models are derived. The conventional inputs are used in the first and second modified RBFNN models (models 3 and 4) whereas exponential nonlinear inputs are used in the fifth and sixth RBFNN models to provide additional nonlinearity for achieving a better solution of nonlinear classification, and direct and inverse modeling problems. To assess and compare the performance potentiality of the proposed four new RBFNN models, one classification problem, one direct modeling problem, and one inverse modeling problem are solved through computer simulation-based experiments. For comparison and to assign the performance rank of each of the four modified RBFNN models, two conventional and commonly used RBFNN models (models 1 and 2) are also simulated. To access the performance of different models during the training phase of Examples 1 and 2, the root mean-square error (RMSE) value, mean absolute deviation (MAD), and the number of iterations required to achieve convergence are obtained. For the third example, only the first two performance measures are found. During the testing or validation phase, the output responses of the different models of Example 2 are compared with the desired response analysis. For Example 3, the bit-error rate (BER) plots are compared. The observation of all the results demonstrates consistent ranks of all models in the case of all three examples. It is, in general, found that the ranks of the models 1–6 are 6, 4, 3, 2, 5, and 1, respectively. In essence, in terms of all performance measures, model M-6 with an exponential version of inputs with weights on both layers occupies the first position whereas model M-4 with conventional inputs as the second position.

An Information Appraisal Procedure: Endows Reliable Online Parameter Identification to Lithium-Ion Battery Model

Online parameter identification is vital for boosting the accuracy of the battery equivalent circuit model (ECM) under dynamic profiles. However, traditional recursive least squares (RLS) method easily decays with the noise corruption from sensors or insufficient exciting signal in reality, which further limits the performance of ECM in battery modeling and states estimation. This article thus proposes a reliable online parameter identification method for battery ECM, which utilizes a well-designed information appraisal procedure based on the Fisher-information-based Cramer–Rao lower bound (CRLB). Without increasing much computing complexity, a comprehensive appraisal indicator, derived recursively from CRLB, enables a new mechanism for online parameter updating. Simulation and experimental results prove the validity of the proposed method under different driving cycles, temperatures, and aging conditions. The results show that the identification accuracy of the proposed method has been significantly improved comparing with a typical RLS and a multiple adaptive forgetting factors RLS method.

Multiwavelength Optoelectronic Synapse with 2D Materials for Mixed-Color Pattern Recognition.

Neuromorphic visual systems emulating biological retina functionalities have enormous potential for in-sensor computing, with prospects of making artificial intelligence ubiquitous. Conventionally, visual information is captured by an image sensor, stored by memory units, and eventually processed by the machine learning algorithm. Here, we present an optoelectronic synapse device with multifunctional integration of all the processes required for real time object identification. Ultraviolet-visible wavelength-sensitive MoS2 FET channel with infrared sensitive PtTe2/Si gate electrode enables the device to sense, store, and process optical data for a wide range of the electromagnetic spectrum, while maintaining a low dark current. The device exhibits optical stimulation-controlled short-term and long-term potentiation, electrically driven long-term depression, synaptic weight update for multiple wavelengths of light ranging from 300 nm in ultraviolet to 2 μm in infrared. An artificial neural network developed using the extracted weight update parameters of the device can be trained to identify both single wavelength and mixed wavelength patterns. This work demonstrates a device that could potentially be used for realizing a multiwavelength neuromorphic visual system for pattern recognition and object identification.

Modified U-Net based photovoltaic array extraction from complex scene in aerial infrared thermal imagery

Hotspot detection in infrared thermal (IRT) images obtained by unmanned aerial vehicles (UAVs) is essential for the reliability, safety and efficiency of photovoltaic (PV) plants. We observe that some objects in the complicated environment, which are outside the region of PV arrays, might significantly interfere with the performance of automatic hotspot detection. This work focuses on the problem of PV array extraction from complex scene, which has not been solved in previous literature. Existing studies have attempted image processing methods to extract PV arrays. However, these approaches rely on manually designed features and parameters, which are impractical for complex scene. In this paper, a deep learning model, namely modified U-Net, is proposed to solve the problem. The proposed model can learn features and parameters automatically. Base on the classical U-Net, several innovations are provided: (1) batch normalization layers are adopted to alleviate internal covariate shift problems; (2) “He initialization” is used to increase the robustness and speed up the convergence; and (3) “RMSprop” is adopted to update parameters adaptively. 1211 IRT images of a PV plant containing complex background are collected to verify the effectiveness of the proposed method. Extensive experiments are conducted to compare the modified U-Net with five deep learning models and three image processing methods. The results demonstrate that the proposed method performs better than existing methods. Cooperating with the modified U-Net for PV array extraction from complex scene, even a simple hotspot detection method can achieve the accuracy of 99.79% and the F1 score of 0.9548.

Integrating Process Mining with Discrete-Event Simulation for Dynamic Productivity Estimation in Heavy Civil Construction Operations

Construction companies are increasingly utilizing sensing technologies to automatically record different steps of the construction process in detail for effective monitoring and control. This generates a significant amount of event data that can be used to learn the underlying behavior of agents in a construction site using process mining. While process mining can be used to discover the real process and identify and analyze deviations and bottlenecks in operations, it is a backward-looking approach. On the other hand, discrete event simulation (DES) provides a means to forecast future performance from historical data to enable proactive decision-making by operation managers relating to their projects. However, this method is largely unused by the industry due to the specialized knowledge required to create the DES models. This paper thus proposes a framework that extends the utility of collecting event data and their process models, by transforming them into DES models for forecasting future performance. This framework also addresses another challenge of using DES relating to its inability to update itself as the project progresses. This challenge is addressed by using the Bayesian updating technique to continuously update the input parameters of the simulation model for the most up-to-date estimation based on data collected from the field. The proposed framework was validated on a real-world case study of an earthmoving operation. The results show that the process mining techniques could accurately discover the process model from the event data collected from the field. Furthermore, it was noted that continuous updating of DES model input parameters can provide accurate and reliable productivity estimates based on the actual data generated from the field. The proposed framework can help stakeholders to discover the underlying sequence of their operations, and enable timely, data-driven decisions regarding operations control.

Improved radial basis function artificial neural network and exact-time extended state observer based non-singular rapid terminal sliding-mode control of quadrotor system

PurposeThe purpose of this paper is to design a hybrid robust tracking controller based on an improved radial basis function artificial neural network (IRBFANN) and a novel extended-state observer for a quadrotor system with various model and parametric uncertainties and external disturbances to enhance the resiliency of the control system.Design/methodology/approachAn IRBFANN is introduced as an adaptive compensator tool for model and parametric uncertainties in the control algorithm of non-singular rapid terminal sliding-mode control (NRTSMC). An exact-time extended state observer (ETESO) augmented with NRTSMC is designed to estimate the unknown exogenous disturbances and ensure fast states convergence while overcoming the singularity issue. The novelty of this work lies in the online updating of weight parameters of the RBFANN algorithm by using a new idea of incorporating an exponential sliding-mode effect, which makes a remarkable effort to make the control protocol adaptive to uncertain model parameters. A comparison of the proposed scheme with other conventional schemes shows its much better performance in the presence of parametric uncertainties and exogenous disturbances.FindingsThe investigated control strategy presents a robust adaptive law based on IRBFANN with a fast convergence rate and improved estimation accuracy via a novel ETESO.Practical implicationsTo enhance the safety level and ensure stable flight operations by the quadrotor in the presence of high-order complex disturbances and uncertain environments, it is imperative to devise a robust control law.Originality/valueA new idea of incorporating an exponential sliding-mode effect instead of conventional approaches in the algorithm of the RBFANN is used, which makes the control law resistant to model and parametric uncertainties. The ETESO provides rapid and accurate disturbance estimation results and updates the control law to overcome the performance degradation caused by the disturbances. Simulation results depict the effectiveness of the proposed control strategy.

Adaptive Trading Strategies Across Liquidity Pools

In this paper, we provide a flexible framework for optimal trading in an asset listed on different venues. We take into account the dependencies between the imbalance and spread of the venues, and allow for partial execution of limit orders at different limits as well as market orders. We present a Bayesian update of the model parameters to take into account possibly changing market conditions and propose extensions to include short/long trading signals, market impact or hidden liquidity. To solve the stochastic control problem of the trader we apply the finite difference method and also develop a deep reinforcement learning algorithm allowing to consider more complex settings.