Spatio-Temporal Event Forecasting Using Incremental Multi-Source Feature Learning

2022 ◽  
Vol 16 (2) ◽  
pp. 1-28
Author(s):  
Liang Zhao ◽  
Yuyang Gao ◽  
Jieping Ye ◽  
Feng Chen ◽  
Yanfang Ye ◽  
...  
Keyword(s):  
Real Time ◽  
Missing Values ◽  
Learning Algorithm ◽  
Feature Learning ◽  
Learning Model ◽  
Model Parameters ◽  
Active Set ◽  
Content Type ◽  
Source Data ◽  
Event Forecasting

The forecasting of significant societal events such as civil unrest and economic crisis is an interesting and challenging problem which requires both timeliness, precision, and comprehensiveness. Significant societal events are influenced and indicated jointly by multiple aspects of a society, including its economics, politics, and culture. Traditional forecasting methods based on a single data source find it hard to cover all these aspects comprehensively, thus limiting model performance. Multi-source event forecasting has proven promising but still suffers from several challenges, including (1) geographical hierarchies in multi-source data features, (2) hierarchical missing values, (3) characterization of structured feature sparsity, and (4) difficulty in model’s online update with incomplete multiple sources. This article proposes a novel feature learning model that concurrently addresses all the above challenges. Specifically, given multi-source data from different geographical levels, we design a new forecasting model by characterizing the lower-level features’ dependence on higher-level features. To handle the correlations amidst structured feature sets and deal with missing values among the coupled features, we propose a novel feature learning model based on an N th-order strong hierarchy and fused-overlapping group Lasso. An efficient algorithm is developed to optimize model parameters and ensure global optima. More importantly, to enable the model update in real time, the online learning algorithm is formulated and active set techniques are leveraged to resolve the crucial challenge when new patterns of missing features appear in real time. Extensive experiments on 10 datasets in different domains demonstrate the effectiveness and efficiency of the proposed models.

Back-propagation of the Mahalanobis distance through a deep triplet learning model for person Re-Identification

2021 ◽  
pp. 1-18
Author(s):  
María José Gómez-Silva ◽  
sArturo de la Escalera ◽  
José María Armingol
Keyword(s):  
Mahalanobis Distance ◽  
Learning Algorithm ◽  
Back Propagation ◽  
Feature Learning ◽  
Learning Model ◽  
Vast Number ◽  
Significant Challenge ◽  
Similar Appearance ◽  
Model Features ◽  
Identification Model

The automatization of the Re-Identification of an individual across different video-surveillance cameras poses a significant challenge due to the presence of a vast number of potential candidates with a similar appearance. This task requires the learning of discriminative features from person images and a distance metric to properly compare them and decide whether they belong to the same person or not. Nevertheless, the fact of acquiring images of the same person from different, distant and non-overlapping views produces changes in illumination, perspective, background, resolution and scale between the person’s representations, resulting in appearance variations that hamper his/her re-identification. This article focuses the feature learning on automatically finding discriminative descriptors able to reflect the dissimilarities mainly due to the changes in actual people appearance, independently from the variations introduced by the acquisition point. With that purpose, such variations have been implicitly embedded by the Mahalanobis distance. This article presents a learning algorithm to jointly model features and the Mahalanobis distance through a Deep Neural Re-Identification model. The Mahalanobis distance learning has been implemented as a novel neural layer, forming part of a Triplet Learning model that has been evaluated over PRID2011 dataset, providing satisfactory results.

Calibration of magnetic compass using an improved extreme learning machine based on reverse tuning

Sensor Review ◽  
2019 ◽  
Vol 39 (1) ◽  
pp. 121-128 ◽  
Author(s):  
Yanxia Liu ◽  
JianJun Fang ◽  
Gang Shi
Keyword(s):  
Real Time ◽  
Bp Neural Network ◽  
Measurement Errors ◽  
Learning Algorithm ◽  
Magnetic Compass ◽  
Content Type ◽  
The Real ◽  
Multi Level ◽  
Learning Machine ◽  
Number Of Iterations

PurposeThe sources of magnetic sensors errors are numerous, such as currents around, soft magnetic and hard magnetic materials and so on. The traditional methods mainly use explicit error models, and it is difficult to include all interference factors. This paper aims to present an implicit error model and studies its high-precision training method.Design/methodology/approachA multi-level extreme learning machine based on reverse tuning (MR-ELM) is presented to compensate for magnetic compass measurement errors by increasing the depth of the network. To ensure the real-time performance of the algorithm, the network structure is fixed to two ELM levels, and the maximum number of levels and neurons will not be continuously increased. The parameters of MR-ELM are further modified by reverse tuning to ensure network accuracy. Because the parameters of the network have been basically determined by least squares, the number of iterations is far less than that in the traditional BP neural network, and the real-time can still be guaranteed.FindingsThe results show that the training time of the MR-ELM is 19.65 s, which is about four times that of the fixed extreme learning algorithm, but training accuracy and generalization performance of the error model are better. The heading error is reduced from the pre-compensation ±2.5° to ±0.125°, and the root mean square error is 0.055°, which is about 0.46 times that of the fixed extreme learning algorithm.Originality/valueMR-ELM is presented to compensate for magnetic compass measurement errors by increasing the depth of the network. In this case, the multi-level ELM network parameters are further modified by reverse tuning to ensure network accuracy. Because the parameters of the network have been basically determined by least squares, the number of iterations is far less than that in the traditional BP neural network, and the real-time training can still be guaranteed. The revised manuscript improved the ELM algorithm itself (referred to as MR-ELM) and bring new ideas to the peers in the magnetic compass error compensation field.

Erroneous branch parameters detection and correction in real time simulation of power systems

2016 ◽  
Vol 35 (5) ◽  
pp. 1656-1671
Author(s):  
Amin Helmzadeh ◽  
Shahram M. Kouhsari
Keyword(s):  
Simulation Model ◽  
State Estimation ◽  
Real Time ◽  
Optimization Procedure ◽  
Optimization Method ◽  
Estimation Methods ◽  
Model Parameters ◽  
Content Type ◽  
Simulation Results ◽  
Correlated Parameters

Purpose The purpose of this paper is to propose an efficient method for detection and modification of erroneous branch parameters in real time power system simulators. The aim of the proposed method is to minimize the sum of squared errors (SSE) due to mismatches between simulation results and corresponding field measurements. Assuming that the network configuration is known, a limited number of erroneous branch parameters will be detected and corrected in an optimization procedure. Design/methodology/approach Proposing a novel formulation that utilizes network voltages and last modified admittance matrix of the simulation model, suspected branch parameters are identified. These parameters are more likely to be responsible for large values of SSE. Utilizing a Gauss-Newton (GN) optimization method, detected parameters will be modified in order to minimize the value of SSE. Required sensitivities in optimization procedure will be calculated numerically by the real time simulator. In addition, by implementing an efficient orthogonalization method, the more effective parameter will be selected among a set of correlated parameters to avoid singularity problems. Findings Unlike state estimation-based methods, the proposed method does not need the mathematical functions of measurements to simulation model parameters. The method can enhance other parameter estimation methods that are based on state estimation. Simulation results demonstrate the high efficiency of the proposed optimization method. Originality/value Incorrect branch parameter detection and correction procedures are investigated in real time simulators.

A Deep Learning Model to Intelligently Identify the Working Status of Screw Pumps for Oil Well Lifting

2021 ◽  
Author(s):  
Zhen Wang ◽  
Yeliang Dong ◽  
Xin Zheng ◽  
Xiang Wang ◽  
Peng Gao ◽  
...  
Keyword(s):  
Deep Learning ◽  
Electric Current ◽  
Real Time ◽  
Learning Model ◽  
Current Data ◽  
Model Parameters ◽  
Screw Pump ◽  
Current Time ◽  
Working Status ◽  
Deep Learning Model

Abstract Screw pumps have been widely used in many oilfields to lift the oil from wellbore to ground. The pump failure and delayed repair means well shut and production loss. A deep learning model is constructed to quickly identify the working status and accurately diagnose the failure types of the screw pumps, which can help the workers always get the information and give a fast repair. Firstly, running parameters of the screw pump, such as electric current, voltage, and instantaneous rate of flow, are obtained through the Real-time Data Acquisition System. Then the correlations between values or trends of those parameters and working status of the screw pump are calculated or analyzed. Results show that there is a good correlation between the current characteristics and various working status of screw pump. Current data at different times are expressed in polar coordinates, with the polar diameter representing the current value and the polar angle representing the time. The current-time curves of massive oil wells are then plotted in images with fixed resolution and divided into nine different groups to correspond to nine frequent working status of screw pump. A convolutional neural network (CNN) model is initialized, with the current-time curve as its input and the number codes representing working status as its output. Images mentioned above are used to train the CNN model, and the model parameters, such as the number of convolution layers, the size of convolution kernels and the activation function are optimized to minimize the training losses, which are the differences between the output codes and the right codes corresponding to the images. Finally, a robust CNN model is established, which can quickly and accurately judge the working state of the screw pump through electric current data. Based on this model, a software system connected with the oilfield database is developed, which can obtain the running parameters of the screw pumps in real time, identify their working states, judge the fault types of the abnormal situations, give alarms, and put forward solution suggestions. The system has now been widely used in Shengli Oilfield, which can help staff know the working conditions and fault types of abnormal wells in real time, speed up the maintenance progress, shorten the pump shutdown time and improve the production.

scholarly journals Road Artery Traffic Light Optimization with Use of the Reinforcement Learning

2014 ◽  
Vol 26 (2) ◽  
pp. 101-108 ◽  
Author(s):  
Rok Marsetič ◽  
Darja Šemrov ◽  
Marijan Žura
Keyword(s):  
Real Time ◽  
Learning Algorithm ◽  
Heuristic Method ◽  
Model Parameters ◽  
Traffic Light ◽  
Q Learning ◽  
Traffic Lights ◽  
Algorithm Comparison ◽  
Actuated Signal ◽  
The Impact

The basic principle of optimal traffic control is the appropriate real-time response to dynamic traffic flow changes. Signal plan efficiency depends on a large number of input parameters. An actuated signal system can adjust very well to traffic conditions, but cannot fully adjust to stochastic traffic volume oscillation. Due to the complexity of the problem analytical methods are not applicable for use in real time, therefore the purpose of this paper is to introduce heuristic method suitable for traffic light optimization in real time. With the evolution of artificial intelligence new possibilities for solving complex problems have been introduced. The goal of this paper is to demonstrate that the use of the Q learning algorithm for traffic lights optimization is suitable. The Q learning algorithm was verified on a road artery with three intersections. For estimation of the effectiveness and efficiency of the proposed algorithm comparison with an actuated signal plan was carried out. The results (average delay per vehicle and the number of vehicles that left road network) show that Q learning algorithm outperforms the actuated signal controllers. The proposed algorithm converges to the minimal delay per vehicle regardless of the stochastic nature of traffic. In this research the impact of the model parameters (learning rate, exploration rate, influence of communication between agents and reward type) on algorithm effectiveness were analysed as well.

Using AI to predict service agent stress from emotion patterns in service interactions

2020 ◽  
Vol ahead-of-print (ahead-of-print) ◽  
Author(s):  
Stefano Bromuri ◽  
Alexander P. Henkel ◽  
Deniz Iren ◽  
Visara Urovi
Keyword(s):  
Deep Learning ◽  
Real Time ◽  
Call Center ◽  
Early Warning Systems ◽  
Learning Model ◽  
Classification Model ◽  
Time Service ◽  
Content Type ◽  
Deep Learning Model ◽  
Service Interactions

PurposeA vast body of literature has documented the negative consequences of stress on employee performance and well-being. These deleterious effects are particularly pronounced for service agents who need to constantly endure and manage customer emotions. The purpose of this paper is to introduce and describe a deep learning model to predict in real-time service agent stress from emotion patterns in voice-to-voice service interactions.Design/methodology/approachA deep learning model was developed to identify emotion patterns in call center interactions based on 363 recorded service interactions, subdivided in 27,889 manually expert-labeled three-second audio snippets. In a second step, the deep learning model was deployed in a call center for a period of one month to be further trained by the data collected from 40 service agents in another 4,672 service interactions.FindingsThe deep learning emotion classifier reached a balanced accuracy of 68% in predicting discrete emotions in service interactions. Integrating this model in a binary classification model, it was able to predict service agent stress with a balanced accuracy of 80%.Practical implicationsService managers can benefit from employing the deep learning model to continuously and unobtrusively monitor the stress level of their service agents with numerous practical applications, including real-time early warning systems for service agents, customized training and automatically linking stress to customer-related outcomes.Originality/valueThe present study is the first to document an artificial intelligence (AI)-based model that is able to identify emotions in natural (i.e. nonstaged) interactions. It is further a pioneer in developing a smart emotion-based stress measure for service agents. Finally, the study contributes to the literature on the role of emotions in service interactions and employee stress.

Parameter adaptation of simplified switched reluctance motor model using Newton and Gauss-Newton signal fitting methods

2017 ◽  
Vol 36 (3) ◽  
pp. 602-618
Author(s):  
Bogdan Fabianski ◽  
Krzysztof Zawirski
Keyword(s):  
Real Time ◽  
Reference Model ◽  
Control Object ◽  
Model Parameters ◽  
Adaptation Process ◽  
Parameter Adaptation ◽  
Content Type ◽  
Switched Reluctance ◽  
Approximation Function ◽  
Model Equations

Purpose The paper is concerned about parameter adaptation of a novel, simplified and nonlinear switched reluctance motor (SRM) model. The purpose of the presented on-line procedure is to give an opportunity to set the model parameters’ values to obtain a relatively good convergence with the real control object. This is important when a reference model is used for control (e.g. optimal) or object state classification (e.g. fault detection) purposes. The more convergent the real object model is, the better operation quality may be expected. Design/methodology/approach In the paper, a 12/8 pole’s SRM as a control object is analyzed. The model equations were verified experimentally by comparing phase current model estimations with reference (measured) ones at different operational points. Differential equations of motor winding currents were chosen as an approximation function in the fitting (parameter adaptation) process using the Newton and Gauss–Newton methods. The structure of the adaptation system is presented along with the implementation in simulation environment. Findings It was confirmed in the simulation tests that Newton and Gauss–Newton methods of nonlinear model parameters’ adaptation may be used for the SRM. The introduced fitting structure is well suited for implementation in real-time, embedded systems. The proposed approximation function could be used in process as an expansion to Jacobian and Hessian matrices. The χ2 (chi2) coefficient (commonly used to measure the quality of the signal fitting) reduced to a low value during the adaptation process. Another introduced quality coefficient shows that the Newton method is slightly better in scope of the entire adaptation process time; however, it needs more computational power. Research limitations/implications The proposed structure and approximation function formula in the parameters’ adaptation system is appropriate for sinusoidal distribution of the motor phase inductance value along the rotor angle position. The inductance angular shape is an implication of the mechanical construction – with appropriate dimensions and materials used. In the presented case, the referenced model is a three-phase SRM in 12/8 poles configuration used as a main drive part of Maytag Neptune washing machine produced by Emerson Motors. Practical implications The presented method of parameter adaptation for novel, simplified and nonlinear SRM model provides an opportunity for its use in embedded, real-time control systems. The convergent motor model, after the fitting procedure (when the estimations are close to the measurements from real object), may be used for solving many well-known control challenges such as detection of initial rotor position, sensorless control, optimal control, fault-tolerant control end in fault detection (FD) systems. The reference model may be used in FD in the way of deducing signals from the difference between the estimated and measured ones. Originality/value The paper proposed a new system of parameter adaptation for the evaluated nonlinear, simplified 12/8 poles SRM model. The relative simplicity of the proposed model equations provides the possibility of implementing an adaptation system in an embedded system that works in a real-time regime. A Two adaptation methods – Newton and Gauss–Newton – have been compared. The obtained results shown that the Newton fitting method is better in the way of the used quality indicator, but it consumes more computational power.

The path of support-to-sales: mediating role of seller collaborative information exchange in social commerce

2020 ◽  
Vol ahead-of-print (ahead-of-print) ◽  
Author(s):  
Li Chen ◽  
Fengxia Zhu ◽  
Murali Mantrala
Keyword(s):  
Information Exchange ◽  
Instrumental Support ◽  
Trade Volume ◽  
Social Commerce ◽  
Content Type ◽  
Mediating Role ◽  
The Social ◽  
Source Data ◽  
Negative Effect

Purpose This paper aims to systematically investigate the direct and indirect effects of four types of support – peer instrumental support, peer emotional support, platform business support and platform communication support – on seller trade volume in social commerce. It also aims to uncover the path of support-to-sales of the seller from a platform perspective and provides a more complete picture of the social commerce phenomenon. Design/methodology/approach This paper uses multi-source data including primary survey data and secondary data on trade volume to test the hypotheses. PROCESS mediation model is used to analyze the multi-source data set. Findings This study finds that the positive effects of peer instrumental support, platform business support and platform communication support on seller trade volume are fully mediated by seller collaborative information exchange. Also, peer emotional support has a significant negative effect on seller trade volume and collaborative information exchange can serve as a buffer to mitigate the negative effect. Research limitations/implications The authors provide new insights into what types of support are or are not conducive to improving transaction volume of individual sellers and highlight the mediating role of seller information exchange in this value generation process in social commerce. These findings advance current knowledge of how seller interactions increase value in social commerce. The chosen research setting may limit the generalizability of the findings of this study. Practical implications This paper offers valuable implications for social commerce platforms on how to better serve their sellers to achieve high growth. Specifically, the findings suggest that platforms should encourage instrumental support and information exchange among peer sellers. In addition, platforms should expand seller support from a single-focus on sellers’ business to a dual-focus on both sellers’ business and socialization in social commerce. Originality/value This paper fulfills an identified need to study how sellers can better derive value from the social interactions and how social commerce platforms can effectively influence transactions, support sales and serve as a selling platform.

scholarly journals Real-time automatic uncertainty estimation of coseismic single rectangular fault model using GNSS data

2021 ◽  
Vol 73 (1) ◽  
Author(s):  
Keitaro Ohno ◽  
Yusaku Ohta ◽  
Satoshi Kawamoto ◽  
Satoshi Abe ◽  
Ryota Hino ◽  
...  
Keyword(s):  
Real Time ◽  
Disaster Response ◽  
Order Approximation ◽  
Nonlinear Problems ◽  
Large Earthquake ◽  
Fault Model ◽  
Deformation Monitoring ◽  
Model Parameters ◽  
The Sea Of Japan ◽  
Plate Interface

AbstractRapid estimation of the coseismic fault model for medium-to-large-sized earthquakes is key for disaster response. To estimate the coseismic fault model for large earthquakes, the Geospatial Information Authority of Japan and Tohoku University have jointly developed a real-time GEONET analysis system for rapid deformation monitoring (REGARD). REGARD can estimate the single rectangular fault model and slip distribution along the assumed plate interface. The single rectangular fault model is useful as a first-order approximation of a medium-to-large earthquake. However, in its estimation, it is difficult to obtain accurate results for model parameters due to the strong effect of initial values. To solve this problem, this study proposes a new method to estimate the coseismic fault model and model uncertainties in real time based on the Bayesian inversion approach using the Markov Chain Monte Carlo (MCMC) method. The MCMC approach is computationally expensive and hyperparameters should be defined in advance via trial and error. The sampling efficiency was improved using a parallel tempering method, and an automatic definition method for hyperparameters was developed for real-time use. The calculation time was within 30 s for 1 × 106 samples using a typical single LINUX server, which can implement real-time analysis, similar to REGARD. The reliability of the developed method was evaluated using data from recent earthquakes (2016 Kumamoto and 2019 Yamagata-Oki earthquakes). Simulations of the earthquakes in the Sea of Japan were also conducted exhaustively. The results showed an advantage over the maximum likelihood approach with a priori information, which has initial value dependence in nonlinear problems. In terms of application to data with a small signal-to-noise ratio, the results suggest the possibility of using several conjugate fault models. There is a tradeoff between the fault area and slip amount, especially for offshore earthquakes, which means that quantification of the uncertainty enables us to evaluate the reliability of the fault model estimation results in real time.

Sign in / Sign up

Export Citation Format

Share Document