Multimodal Data-Driven Reinforcement Learning for Operational Decision-Making in Industrial Processes

Dear Editor, This letter proposes a multimodal data-driven reinforcement learning-based method for operational decision-making in industrial processes. Due to the frequent fluctuations of feedstock properties and operating conditions in the industrial processes, existing data-driven methods cannot effectively adjust the operational variables. In addition, multimodal data such as images, audio.

Multi-Perspective Data Fusion Framework Based on Hierarchical BERT: Provide Visual Predictions of Business Processes

Predictive Business Process Monitoring(PBPM)is a significant research area in Business Process Management(BPM)aimed at accurately forecasting future behavioral events.At present,deep learning methods are widely cited in PBPM research,but no method has been effective in fusing data information into the control flow for multi-perspective process prediction.Therefore,this paper proposes a process prediction method based on the hierarchical BERT and multi-perspective data fusion.Firstly,the first layer BERT network learns the correlations between different category attribute data.Then,the attribute data is integrated into a weighted event-level feature vector and input into the second layer BERT network to learn the impact and priority relationship of each event on future predicted events.Next,the multi-head attention mechanism within the framework is visualized for analysis,helping to understand the decision-making logic of the framework and providing visual predictions.Finally,experimental results show that the predictive accuracy of the framework surpasses the current state-of-the-art research methods and significantly enhances the predictive performance of BPM.

A novel method for clustering cellular data to improve classification

Many fields,such as neuroscience,are experiencing the vast prolife ration of cellular data,underscoring the need fo r organizing and interpreting large datasets.A popular approach partitions data into manageable subsets via hierarchical clustering,but objective methods to determine the appropriate classification granularity are missing.We recently introduced a technique to systematically identify when to stop subdividing clusters based on the fundamental principle that cells must differ more between than within clusters.Here we present the corresponding protocol to classify cellular datasets by combining datadriven unsupervised hierarchical clustering with statistical testing.These general-purpose functions are applicable to any cellular dataset that can be organized as two-dimensional matrices of numerical values,including molecula r,physiological,and anatomical datasets.We demonstrate the protocol using cellular data from the Janelia MouseLight project to chara cterize morphological aspects of neurons.

Synthetic data as an investigative tool in hypertension and renal diseases research

There is a growing body of clinical research on the utility of synthetic data derivatives,an emerging research tool in medicine.In nephrology,clinicians can use machine learning and artificial intelligence as powerful aids in their clinical decision-making while also preserving patient privacy.This is especially important given the epidemiology of chronic kidney disease,renal oncology,and hypertension worldwide.However,there remains a need to create a framework for guidance regarding how to better utilize synthetic data as a practical application in this research.

Microphysical Processes of a Stratiform Precipitation Event over Eastern China:Analysis Using Micro Rain Radar data

Data collected using the micro rain radar(MRR) situated in Jinan city, eastern China, were used to explore the altitudinal and temporal evolution of rainfall microphysical characteristics, and to analyze the bright band(BB) characteristics and hydrometeor classification. Specifically, a low-intensity and stable stratiform precipitation event that occurred from 0000 to0550 UTC 15 February 2015 and featured a BB was studied. During this event, the rainfall intensity was less than 2 mm h-1 at a height of 300 m, which was above the radar site level, so the errors caused by the vertical air motion could be ignored.The freezing height from the radiosonde matched well with the top of the BB observed by the MRR. It was also found that the number of 0.5–1 mm diameter drops showed no noticeable variation below the BB. The maximum fall velocity and the maximum gradient fall velocity(GFV) of the raindrops appeared at the bottom of the BB. Meanwhile, a method that uses the GFV and reflectivity to identify the altitude and the thickness of the BB was established, with which the MRR can provide a reliable and real-time estimation of the 0?C isotherm. The droplet fall velocity was used to classify the types of snow crystals above the BB. In the first 20 min of the selected precipitation event, graupel prevailed above the BB; and at an altitude of2000 m, graupel also dominated in the first 250 min. After 150 min, the existence of graupel and dendritic crystals with water droplets above the BB was inferred.

Application of Radar-Measured Rain Data in Hydrological Processes Modeling during the Intensified Observation Period of HUBEX

On the basis of Digital Elevation Model data, the raster flow vectors, watershed delineation, and spatial topological relationship are generated by the Martz and Garbrecht method for the upper area of Huangnizhuang station in the Shihe Catchment with 805 km2 of area, an intensified observation field for the HUBEX/GAME Project. Then, the Xin’anjiang Model is applied for runoff production in each grid element where rain data measured by radar at Fuyang station is utilized as the input of the hydrological model. The elements are connected by flow vectors to the outlet of the drainage catchment where runoff is routed by the Muskingum method from each grid element to the outlet according to the length between each grid and the outlet. The Nash-Sutcliffe model efficiency coefficient is 92.41% from 31 May to 3 August 1998, and 85.64%, 86.62%, 92.57%, and 83.91%, respectively for the 1st, 2nd, 3rd, and 4th flood events during the whole computational period. As compared with the case where rain-gauge data are used in simulating the hourly hydrograph at Huangnizhuang station in the Shihe Catchment, the index of model efficiency improvement is positive, ranging from 27.56% to 69.39%. This justifies the claim that radar-measured data are superior to rain-gauge data as inputs to hydrological modeling. As a result, the grid-based hydrological model provides a good platform for runoff computation when radar-measured rain data with highly spatiotemporal resolution are taken as the input of the hydrological model.

基于re3data的中英科学数据仓储平台对比研究

以re3data为数据获取源,选取中英两国406个科学数据仓储为研究对象,从分布特征、责任类型、仓储许可、技术标准及质量标准等5个方面、11个指标对两国科学数据仓储的建设情况进行对比分析,试图为我国数据仓储的可持续发展提出建议:广泛联结国内外异质机构,推进多学科领域的交流与合作,有效扩充仓储许可权限与类型,优化技术标准的应用现况,提高元数据使用的灵活性。

Baddeleyite from Large Complex Deposits: Significance for Archean-Paleozoic Plume Processes in the Arctic Region (NE Fennoscandian Shield) Based on U-Pb (ID-TIMS) and LA-ICP-MS Data

Baddeleyite is an important mineral geochronometer. It is valued in the U-Pb (ID-TIMS) geochronology more than zircon because of its magmatic origin, while zircon can be metamorphic, hydrothermal or occur as xenocrysts. Detailed mineralogical (BSE, KL, etc.) research of baddeleyite started in the Fennoscandian Shield in the 1990s. The mineral was first extracted from the Paleozoic Kovdor deposit, the second-biggest baddeleyite deposit in the world after Phalaborwa (2.1 Ga), South Africa. The mineral was successfully introduced into the U-Pb systematics. This study provides new U-Pb and LA-ICP-MS data on Archean Ti-Mgt and BIF deposits, Paleoproterozoic layered PGE intrusions with Pt-Pd and Cu-Ni reefs and Paleozoic complex deposits (baddeleyite, apatite, foscorite ores, etc.) in the NE Fennoscandian Shield. Data on concentrations of REE in baddeleyite and temperature of the U-Pb systematics closure are also provided. It is shown that baddeleyite plays an important role in the geological history of the Earth, in particular, in the break-up of supercontinents.

Analysis of sea level changes in the Caspian Sea related to Cosmo-geophysical processes based on satellite and terrestrial data

Analysis results of the average annual sea levels in the Caspian Sea obtained from ground and satellite observations, corresponding to solar activity characteristics, magnetic field data, and length of day are presented. Spectra of the indicated processes were investigated and their approximation models were also built. Previously assumed statistical relationships between space-geophysical processes and Caspian Sea level(CSL) changes were confirmed. A close connection was revealed between the low-frequency models of the solar and geomagnetic activity parameters and the CSL changes. Predictions extending into the next decades showed a high probability of an increase in the CSL and a decrease of the compared space-geophysical parameters.

Search Processes in the Exploration of Complex Data under Different Display Conditions

The study investigated user experience, display complexity, display type (tables versus graphs), and task difficulty as variables affecting the user’s ability to navigate through complex visual data. A total of 64 participants, 39 undergraduate students (novice users) and 25 graduate students (intermediate-level users) participated in the study. The experimental design was 2 × 2 × 2 × 3 mixed design using two between-subject variables (display complexity, user experience) and two within-subject variables (display format, question difficulty). The results indicated that response time was superior for graphs (relative to tables), especially when the questions were difficult. The intermediate users seemed to adopt more extensive search strategies than novices, as revealed by an analysis of the number of changes they made to the display prior to answering questions. It was concluded that designers of data displays should consider the (a) type of display, (b) difficulty of the task, and (c) expertise level of the user to obtain optimal levels of performance.

Cloud-Edge Collaborative Federated GAN Based Data Processing for IoT-Empowered Multi-Flow Integrated Energy Aggregation Dispatch

The convergence of Internet of Things(IoT),5G,and cloud collaboration offers tailored solutions to the rigorous demands of multi-flow integrated energy aggregation dispatch data processing.While generative adversarial networks(GANs)are instrumental in resource scheduling,their application in this domain is impeded by challenges such as convergence speed,inferior optimality searching capability,and the inability to learn from failed decision making feedbacks.Therefore,a cloud-edge collaborative federated GAN-based communication and computing resource scheduling algorithm with long-term constraint violation sensitiveness is proposed to address these challenges.The proposed algorithm facilitates real-time,energy-efficient data processing by optimizing transmission power control,data migration,and computing resource allocation.It employs federated learning for global parameter aggregation to enhance GAN parameter updating and dynamically adjusts GAN learning rates and global aggregation weights based on energy consumption constraint violations.Simulation results indicate that the proposed algorithm effectively reduces data processing latency,energy consumption,and convergence time.

Inter-agency government information sharing under data-driven blockchain framework

The inter-agency government information sharing(IAGIS)plays an important role in improving service and efficiency of government agencies.Currently,there is still no effective and secure way for data-driven IAGIS to fulfill dynamic demands of information sharing between government agencies.Motivated by blockchain and data mining,a data-driven framework is proposed for IAGIS in this paper.Firstly,the blockchain is used as the core to design the whole framework for monitoring and preventing leakage and abuse of government information,in order to guarantee information security.Secondly,a four-layer architecture is designed for implementing the proposed framework.Thirdly,the classical data mining algorithms PageRank and Apriori are applied to dynamically design smart contracts for information sharing,for the purposed of flexibly adjusting the information sharing strategies according to the practical demands of government agencies for public management and public service.Finally,a case study is presented to illustrate the operation of the proposed framework.

Enhancing the vertical resolution of lunar penetrating radar data using predictive deconvolution

The Yutu-2 rover onboard the Chang’E-4 mission performed the first lunar penetrating radar detection on the farside of the Moon.The high-frequency channel presented us with many unprecedented details of the subsurface structures within a depth of approximately 50 m.However,it was still difficult to identify finer layers from the cluttered reflections and scattering waves.We applied deconvolution to improve the vertical resolution of the radar profile by extending the limited bandwidth associated with the emissive radar pulse.To overcome the challenges arising from the mixed-phase wavelets and the problematic amplification of noise,we performed predictive deconvolution to remove the minimum-phase components from the Chang’E-4 dataset,followed by a comprehensive phase rotation to rectify phase anomalies in the radar image.Subsequently,we implemented irreversible migration filtering to mitigate the noise and diminutive clutter echoes amplified by deconvolution.The processed data showed evident enhancement of the vertical resolution with a widened bandwidth in the frequency domain and better signal clarity in the time domain,providing us with more undisputed details of subsurface structures near the Chang’E-4 landing site.

Hadoop-based secure storage solution for big data in cloud computing environment

In order to address the problems of the single encryption algorithm,such as low encryption efficiency and unreliable metadata for static data storage of big data platforms in the cloud computing environment,we propose a Hadoop based big data secure storage scheme.Firstly,in order to disperse the NameNode service from a single server to multiple servers,we combine HDFS federation and HDFS high-availability mechanisms,and use the Zookeeper distributed coordination mechanism to coordinate each node to achieve dual-channel storage.Then,we improve the ECC encryption algorithm for the encryption of ordinary data,and adopt a homomorphic encryption algorithm to encrypt data that needs to be calculated.To accelerate the encryption,we adopt the dualthread encryption mode.Finally,the HDFS control module is designed to combine the encryption algorithm with the storage model.Experimental results show that the proposed solution solves the problem of a single point of failure of metadata,performs well in terms of metadata reliability,and can realize the fault tolerance of the server.The improved encryption algorithm integrates the dual-channel storage mode,and the encryption storage efficiency improves by 27.6% on average.

Data Secure Storage Mechanism for IIoT Based on Blockchain

With the development of Industry 4.0 and big data technology,the Industrial Internet of Things(IIoT)is hampered by inherent issues such as privacy,security,and fault tolerance,which pose certain challenges to the rapid development of IIoT.Blockchain technology has immutability,decentralization,and autonomy,which can greatly improve the inherent defects of the IIoT.In the traditional blockchain,data is stored in a Merkle tree.As data continues to grow,the scale of proofs used to validate it grows,threatening the efficiency,security,and reliability of blockchain-based IIoT.Accordingly,this paper first analyzes the inefficiency of the traditional blockchain structure in verifying the integrity and correctness of data.To solve this problem,a new Vector Commitment(VC)structure,Partition Vector Commitment(PVC),is proposed by improving the traditional VC structure.Secondly,this paper uses PVC instead of the Merkle tree to store big data generated by IIoT.PVC can improve the efficiency of traditional VC in the process of commitment and opening.Finally,this paper uses PVC to build a blockchain-based IIoT data security storage mechanism and carries out a comparative analysis of experiments.This mechanism can greatly reduce communication loss and maximize the rational use of storage space,which is of great significance for maintaining the security and stability of blockchain-based IIoT.

Data processing method for aerial testing of rotating accelerometer gravity gradiometer

A novel method for noise removal from the rotating accelerometer gravity gradiometer(MAGG)is presented.It introduces a head-to-tail data expansion technique based on the zero-phase filtering principle.A scheme for determining band-pass filter parameters based on signal-to-noise ratio gain,smoothness index,and cross-correlation coefficient is designed using the Chebyshev optimal consistent approximation theory.Additionally,a wavelet denoising evaluation function is constructed,with the dmey wavelet basis function identified as most effective for processing gravity gradient data.The results of hard-in-the-loop simulation and prototype experiments show that the proposed processing method has shown a 14%improvement in the measurement variance of gravity gradient signals,and the measurement accuracy has reached within 4E,compared to other commonly used methods,which verifies that the proposed method effectively removes noise from the gradient signals,improved gravity gradiometry accuracy,and has certain technical insights for high-precision airborne gravity gradiometry.

Defect Detection Model Using Time Series Data Augmentation and Transformation

Time-series data provide important information in many fields,and their processing and analysis have been the focus of much research.However,detecting anomalies is very difficult due to data imbalance,temporal dependence,and noise.Therefore,methodologies for data augmentation and conversion of time series data into images for analysis have been studied.This paper proposes a fault detection model that uses time series data augmentation and transformation to address the problems of data imbalance,temporal dependence,and robustness to noise.The method of data augmentation is set as the addition of noise.It involves adding Gaussian noise,with the noise level set to 0.002,to maximize the generalization performance of the model.In addition,we use the Markov Transition Field(MTF)method to effectively visualize the dynamic transitions of the data while converting the time series data into images.It enables the identification of patterns in time series data and assists in capturing the sequential dependencies of the data.For anomaly detection,the PatchCore model is applied to show excellent performance,and the detected anomaly areas are represented as heat maps.It allows for the detection of anomalies,and by applying an anomaly map to the original image,it is possible to capture the areas where anomalies occur.The performance evaluation shows that both F1-score and Accuracy are high when time series data is converted to images.Additionally,when processed as images rather than as time series data,there was a significant reduction in both the size of the data and the training time.The proposed method can provide an important springboard for research in the field of anomaly detection using time series data.Besides,it helps solve problems such as analyzing complex patterns in data lightweight.

Analysis of Morphological Processes in a Disturbed Gravel-Bed River （Piave River）： Integration of LiDAR Data and Colour Bathymetry

The magnitude of river morphological changes are better analyzed through the use of quantitative approaches, wherein resolution accuracy and uncertainty assessment are treated as crucial key-factors. In this sense, the creation of precise DEMs （Digital Elevation Models） of rivers represents an affordable tool to analyze geomorphic variations and budgets, except for wetted areas, where reliable channel digitalization can normally be obtained only using expensive bathymetric surveys. The proposed work aims at improving channel surface models without having available bathymetric sensors, by deriving dry areas elevations from LiDAR data and water depth of wetted areas from aerial photos through a predictive depth-colour relationship. The methodology was applied to two different sub-reaches of the Piave River, a gravel-bed river which suffered severe flood events in 2010. Erosion and deposition patterns were identified through DEM differencing, showing a predominance of scour processes which can lead to channel instability situations. The bathymetric output was compared to other previously-derived models confirming the accuracy of the in-channel elevation estimates. Finally, a discussion on the role played by longitudinal protections during the studied flood events is proposed, focusing the attention on the incidence of two major bank erosions that removed significant volumes of stable areas.

Constraints on Characteristics and Distribution of Gas Hydrate and Free Gas Using Broad-Band Processing of Three-Dimensional Seismic Data

Gas hydrate drilling expeditions in the Pearl River Mouth Basin,South China Sea,have identified concentrated gas hydrates with variable thickness.Moreover,free gas and the coexistence of gas hydrate and free gas have been confirmed by logging,coring,and production tests in the foraminifera-rich silty sediments with complex bottom-simulating reflectors(BSRs).The broad-band processing is conducted on conventional three-dimensional(3D)seismic data to improve the image and detection accuracy of gas hydratebearing layers and delineate the saturation and thickness of gas hydrate-and free gas-bearing sediments.Several geophysical attributes extracted along the base of the gas hydrate stability zone are used to demonstrate the variable distribution and the controlling factors for the differential enrichment of gas hydrate.The inverted gas hydrate saturation at the production zone is over 40% with a thickness of 90 m,showing the interbedded distribution with different boundaries between gas hydrate-and free gas-bearing layers.However,the gas hydrate saturation value at the adjacent canyon is 70%,with 30-m-thick patches and linear features.The lithological and fault controls on gas hydrate and free gas distributions are demonstrated by tracing each gas hydrate-bearing layer.Moreover,the BSR depths based on broad-band reprocessed 3D seismic data not only exhibit variations due to small-scale topographic changes caused by seafloor sedimentation and erosion but also show the upward shift of BSR and the blocky distribution of the coexistence of gas hydrate and free gas in the Pearl River Mouth Basin.

Enhanced prediction of anisotropic deformation behavior using machine learning with data augmentation

Mg alloys possess an inherent plastic anisotropy owing to the selective activation of deformation mechanisms depending on the loading condition.This characteristic results in a diverse range of flow curves that vary with a deformation condition.This study proposes a novel approach for accurately predicting an anisotropic deformation behavior of wrought Mg alloys using machine learning(ML)with data augmentation.The developed model combines four key strategies from data science:learning the entire flow curves,generative adversarial networks(GAN),algorithm-driven hyperparameter tuning,and gated recurrent unit(GRU)architecture.The proposed model,namely GAN-aided GRU,was extensively evaluated for various predictive scenarios,such as interpolation,extrapolation,and a limited dataset size.The model exhibited significant predictability and improved generalizability for estimating the anisotropic compressive behavior of ZK60 Mg alloys under 11 annealing conditions and for three loading directions.The GAN-aided GRU results were superior to those of previous ML models and constitutive equations.The superior performance was attributed to hyperparameter optimization,GAN-based data augmentation,and the inherent predictivity of the GRU for extrapolation.As a first attempt to employ ML techniques other than artificial neural networks,this study proposes a novel perspective on predicting the anisotropic deformation behaviors of wrought Mg alloys.