Search

In unsupervised domain adaptation (UDA), a classifier for the target domain is trained with massive true-label data from the source domain and unlabeled data from the target domain. However, collecting fully-true-label data in the source domain is high-cost and sometimes impossible. Compared to the true labels, a complementary label specifies a class that a pattern does not belong to, hence collecting complementary labels would be less laborious than collecting true labels. Thus, in this paper, we propose a novel setting that the source domain is composed of complementary-label data, and a theoretical bound for it is first proved. We consider two cases of this setting, one is that the source domain only contains complementary-label data (completely complementary unsupervised domain adaptation, CC-UDA), and the other is that the source domain has plenty of complementary-label data and a small amount of true-label data (partly complementary unsupervised domain adaptation, PC-UDA). To this end, a complementary label adversarial network} (CLARINET) is proposed to solve CC-UDA and PC-UDA problems. CLARINET maintains two deep networks simultaneously, where one focuses on classifying complementary-label source data and the other takes care of source-to-target distributional adaptation. Experiments show that CLARINET significantly outperforms a series of competent baselines on handwritten-digits-recognition and objects-recognition tasks., arXiv admin note: text overlap with arXiv:2007.14612

Domain adaptation aims at extracting knowledge from auxiliary source domains to assist the learning task in a target domain. In classification problems, since the distributions of the source and target domains are different, directly using source data to build a classifier for the target domain may hamper the classification performance on the target data. Fortunately, in many tasks, there can be some features that are transferable, i.e., the source and target domains share similar properties. On the other hand, it is common that the source data contain noisy features which may degrade the learning performance in the target domain. This issue, however, is barely studied in existing works. In this paper, we propose to find a feature subset that is transferable across the source and target domains. As a result, the domain discrepancy measured on the selected features can be reduced. Moreover, we seek to find the most discriminative features for classification. To achieve the above goals, we formulate a new sparse learning model that is able to jointly reduce the domain discrepancy and select informative features for classification. We develop two optimization algorithms to address the derived learning problem. Extensive experiments on real-world data sets demonstrate the effectiveness of the proposed method.

This work introduces a generic, machine learning (ML)-based platform for functional reverse engineering (RE) of circuits. Our proposed platform GNN-RE leverages the notion of graph neural networks (GNNs) to (i) represent and analyze flattened/ unstructured gate-level netlists, (ii) automatically identify the boundaries between the modules or sub-circuits implemented in such netlists and (iii) classify the sub-circuits based on their functionalities. For GNNs in general, each graph node is tailored to learn about its own features and its neighboring nodes, which is a powerful approach for the detection of any kind of sub-graphs of interest. For GNN-RE, in particular, each node represents a gate and is initialized with a feature vector that reflects on the functional and structural properties of its neighboring gates. GNN-RE also learns the global structure of the circuit, which facilitates identifying the boundaries between subcircuits in a flattened netlist. Initially, to provide high-quality data for training of GNN-RE, we deploy a comprehensive dataset of foundational designs/components with differing functionalities, implementation styles, bit-widths, and interconnections. GNN-RE is then tested on the unseen shares of this custom dataset, as well as the EPFL benchmarks, the ISCAS-85 benchmarks, and the 74X series benchmarks. GNN-RE achieves an average accuracy of 98:82% in terms of mapping individual gates to modules, all without any manual intervention or post-processing. We also release our code and source data 1.

Thermal sensors play an important role in intelligent transportation system. This paper studies the problem of RGB and thermal (RGBT) tracking in challenging situations by leveraging multimodal data. A RGBT object tracking method is proposed in correlation filter tracking framework based on short term historical information. Given the initial object bounding box, hierarchical convolutional neural network (CNN) is employed to extract features. The target is tracked for RGB and thermal modalities separately. Then the backward tracking is implemented in the two modalities. The difference between each pair is computed, which is an indicator of the tracking quality in each modality. Considering the temporal continuity of sequence frames, we also incorporate the history data into the weights computation to achieve a robust fusion of different source data. Experiments on three RGBT datasets show the proposed method achieves comparable results to state-of-the-art methods.

Most of the classification methods of urban functional areas nowadays are only based on single source data analysis and modeling, which can not make full use of the multi-scale and multi-source data that is easy to obtain. Therefore, this paper proposed a classification model of urban functional areas based on multi-modal machine learning, by analyzing regional remote sensing images and behavior data of visitors in the area, using the combination of supervised methods extracted the deep-seated features and relationships of kinds of data, filtered and merged the overall and local features of the data. The model used dual branch neural network combining SE-ResNeXt and Dual Path Network (DPN) to automatically mined and fused the overall characteristics of multi-source data, and used the designed feature engineering to deeply mine the behavior data of users to obtain more association information, then combined the algorithm based on Gradient Boosting Decision Tree to learn the characteristics of different levels and obtained the classification probability for different levels of features. Finally, we continued to use the algorithm based on the Gradient Boosting Decision Tree to learn the probability distribution of different levels of features to obtain the final prediction results of urban functional area classification. Through the analysis and experimental verification of real data sets, the results showed that MM-UrbanFAC model can effectively integrate the features of multi-modal data. Compared with a single classifier, the integration framework based on gradient lifting tree improved the prediction performance, this method can effectively integrate the results of multiple models and accurately classify urban functional areas, and the model can provide reference for tourism recommendation, urban land planning and urban construction.

Speech emotion recognition has become an attractive research topic due to various emotional states of speech signals in real-life scenarios. Most current speech emotion recognition methods are carried out on a single corpus. However, in practice, the training and testing data often come from different domains, e.g., different corpora. In this case, the model generalizability and recognition performance would decrease greatly due to the domain mismatch. To address this challenging problem, we present a transfer learning method, called joint transfer subspace learning and regression (JTSLR), for cross-corpus speech emotion recognition. Specifically, JTSLR performs transfer subspace learning and regression in a joint framework. First, we learn a latent subspace by introducing a discriminative maximum mean discrepancy (MMD) as the discrepancy metric. Then, we put forward a regression function in this latent subspace to describe the relationships between features and corresponding labels. Moreover, we present a label graph to help transfer knowledge from relevant source data to target data. Finally, we conduct extensive experiments on three popular emotional datasets. The results show that our method can outperform traditional methods and some state-of-the-art transfer learning algorithms for cross-corpus speech emotion recognition tasks.

Unsupervised domain adaptation (UDA) in person re-identification (re-ID) is a challenging task, aiming to learn a model with labeled source data and unlabeled target data to recognize the same person in the target domain across different cameras. Recently, a lot of popular and promising methods based on clustering are proposed for this task and achieve a sizable progress. However, in those methods, without target labels, the clustering algorithms will inevitably produce noisy pseudo-labels. Overfitting on these noisy labels is severely harmful to the performance and generalization of models. In order to address the above issues, we propose a novel framework, Adaptive Deep Clustering (AdaDC), to reduce the negative impact of noisy pseudo-labels. On one hand, the proposed approach employs different clustering methods adaptively and alternately to fully exploit their complementary information and avoid overfitting noisy pseudo-labels. On the other hand, there is a progressive sample selection strategy for reducing noisy label ratio in pseudo-labels, which is achieved by integrating different clustering results. Experiments present that the proposed approach can achieve state-of-the-art performance compared to the other recent UDA person re-ID methods on widely-used datasets. Moreover, there are some other analysis experiments conducted for verifying the effectiveness of the proposed approach.

In recent years, many applications are using various forms of deep learning models. Such methods are usually based on traditional learning paradigms requiring the consistency of properties among the feature spaces of the training and test data and also the availability of large amounts of training data, e.g., for performing supervised learning tasks. However, many real-world data do not adhere to such assumptions. In such situations transfer learning can provide feasible solutions, e.g., by simultaneously learning from data-rich source data and data-sparse target data to transfer information for learning a target task. In this paper, we survey deep transfer learning models with a focus on applications to text data. First, we review the terminology used in the literature and introduce a new nomenclature allowing the unequivocal description of a transfer learning model. Second, we introduce a visual taxonomy of deep learning approaches that provides a systematic structure to the many diverse models introduced until now. Furthermore, we provide comprehensive information about text data that have been used for studying such models because only by the application of methods to data, performance measures can be estimated and models assessed. publishedVersion

The utility of any database or registry depends on the completeness and accuracy of the data it contains. This report documents the validity of data elements within DataDerm, the clinical registry database of the American Academy of Dermatology. An external audit of DataDerm, performed by a third-party vendor, involved the manual review of 1098 individual patient charts from calendar year 2018 from 8 different dermatology practices that used 4 different electronic health records. At each site, 142 discrete data fields were assessed, comparing the data within DataDerm to the source data within the electronic health record. Audited data included 3 domains of data elements (diagnoses, medications, and procedures) and a performance measure ("Biopsy Reporting Time-Clinician to Patient"), which is 1 of several measures used by DataDerm as a Qualified Clinical Data Registry. Overall completeness of data was 95.3%, with a range among practices of 90.6% to 98.5%. Overall accuracy of data was 89.8%, with a range of accuracy among practices of 81.2% to 94.1%. These levels of completeness and accuracy exceed the rates in the literature for registries that are based on data that is extracted from electronic health records; and therefore, this audit validates the excellent quality of data in DataDerm.

Domain adaptation for object detection has been extensively studied in recent years. Most existing approaches focus on single-source unsupervised domain adaptive object detection. However, a more practical scenario is that the labeled source data is collected from multiple domains with different feature distributions. The conventional approaches do not work very well since multiple domain gaps exist. We propose a Multi-source domain Knowledge Transfer (MKT) method to handle this situation. First, the low-level features from multiple domains are aligned by learning a shallow feature extraction network. Then, the high-level features from each pair of source and target domains are aligned by the followed multi-branch network. After that, we perform two parts of information fusion: (1) We train a detection network shared by all branches based on the transferability of each source sample feature. The transferability of a source sample feature means the indistinguishable degree to the target domain sample features. (2) For using our model, the target sample features output by the multi-branch network are fused based on the average transferability of each domain. Moreover, we leverage both image-level and instance-level attention to promote positive cross-domain transfer and suppress negative transfer. Our main contributions are the two-stage feature alignments and information fusion. Extensive experimental results on various transfer scenarios show that our method achieves the state-of-the-art performance.

Missing data reconstruction is a classical yet challenging problem in remote sensing images. Most current methods based on traditional convolutional neural network require supplementary data and can only handle one specific task. To address these limitations, we propose a novel generative adversarial network-based missing data reconstruction method in this letter, which is capable of various reconstruction tasks given only single source data as input. Two auxiliary patch-based discriminators are deployed to impose additional constraints on the local and global regions, respectively. In order to better fit the nature of remote sensing images, we introduce special convolutions and attention mechanism in a two-stage generator, thereby benefiting the tradeoff between accuracy and efficiency. Combining with perceptual and multiscale adversarial losses, the proposed model can produce coherent structure with better details. Qualitative and quantitative experiments demonstrate the uncompromising performance of the proposed model against multisource methods in generating visually plausible reconstruction results. Moreover, further exploration shows a promising way for the proposed model to utilize spatio-spectral-temporal information. The codes and models are available at https://github.com/Oliiveralien/Inpainting-on-RSI.

Classifying hard samples in the course of RGBT tracking is a quite challenging problem. Existing methods only focus on enlarging the boundary between positive and negative samples, but ignore the relations of multilevel hard samples, which are crucial for the robustness of hard sample classification. To handle this problem, we propose a novel Multi-Modal Multi-Margin Metric Learning framework named M5L for RGBT tracking. In particular, we divided all samples into four parts including normal positive, normal negative, hard positive and hard negative ones, and aim to leverage their relations to improve the robustness of feature embeddings, e.g., normal positive samples are closer to the ground truth than hard positive ones. To this end, we design a multi-modal multi-margin structural loss to preserve the relations of multilevel hard samples in the training stage. In addition, we introduce an attention-based fusion module to achieve quality-aware integration of different source data. Extensive experiments on large-scale datasets testify that our framework clearly improves the tracking performance and performs favorably the state-of-the-art RGBT trackers.

Noting the fact that the high-dimensional data composed of various polarimetric features has better decouplability than polarimetric synthetic aperture radar (PolSAR) image source data, in this letter, multiple polarimetric features are extracted and stacked to form a high-dimensional feature cube as the input of convolutional neural networks (CNNs) to improve the performance of PolSAR image classification. Directly utilizing the polarimetric features will produce a performance degradation and the recalibration of them is indispensable. However, classical feature selection methods are independent of the classifier, which means that the stimulated features may not be the classification-friendly ones. To avoid separated procedures and improve the performance, attention-based polarimetric feature selection convolutional network, called AFS-CNN, is proposed to implement end-to-end feature selection and classification. The relationship between input polarimetric features can be captured and embedded through attention-based architecture to ensure the validity of high-dimensional data classification. Experiments on two PolSAR benchmark data sets verify the performance of the proposed method. Furthermore, this work is quite flexible, which is reflected in that the proposal can be used as a plug-and-play component of any CNN-based PolSAR classifier.

In recent years, with the development of high-resolution remote sensing (RS) images and deep learning technology, high-quality source data and state-of-the-art methods have become increasingly available, and great progress has been made in change detection (CD) in RS fields. However, existing methods still suffer from weak network feature representation and poor CD performance. To address these problems, we propose a novel CD network, called dual-pathway CD network (DP-CD-Net), which can help enhance feature representation and achieve a more accurate difference map. The proposed method contains a dual-pathway feature difference network (FDN), an adaptive fusion module (AFM), and an auxiliary supervision strategy. Dual-pathway FDNs can effectively enhance feature representation by supplementing the detailed information from the encoding layers. Then, we use the AFM method to fuse the difference maps. To solve the problem of training difficulty, we use the auxiliary supervision strategy to improve the performance of DP-CD-Net. We conduct extensive experiments to validate the performance of the proposed method on the LEVIR-CD dataset. The results demonstrate that the proposed method performs better than existing methods.

The simultaneous source acquisition method, which excites multiple sources in a narrow time interval, can greatly improve the efficiency of seismic data acquisition and provide good illumination. However, the simultaneous source data, also known as the blended data, contain the crosstalk noise from other sources, which brings trouble to the subsequent processing flow. Therefore, an effective deblending method for the simultaneous source data is needed. In order to suppress crosstalk noise, an iterative deblending method using a deep neural network (DNN) trained in the common shot gather (CSG) is developed in this paper with the double-blended simultaneous source (DBSS) data being the input data and the blended CSG data being the label data. The proposed training method can not only solve the problem of difficult acquisition of the label data, but also make the DNN applicable to any complex formation conditions without considering whether the DNN has the generalization ability of deblending in different work areas. In the test phase, the trained DNN is embedded into the iterative separation framework to deblend the data in the common receiver gather (CRG), which can achieve convergence in a few iterations and achieve a better separation effect. The synthetic and field data examples are tested to verify that the proposed method can effectively suppress the crosstalk noise when deblending the simultaneous source data.

The ASH Research Collaborative is a nonprofit organization established through the American Society of Hematology’s commitment to patients with hematologic conditions and the science that informs clinical care and future therapies. The ASH Research Collaborative houses 2 major initiatives: (1) the Data Hub and (2) the Clinical Trials Network (CTN). The Data Hub is a program for hematologic diseases in which networks of clinical care delivery sites are developed in specific disease areas, with individual patient data contributed through electronic health record (EHR) integration, direct data entry through electronic data capture, and external data sources. Disease-specific data models are constructed so that data can be assembled into analytic datasets and used to enhance clinical care through dashboards and other mechanisms. Initial models have been built in multiple myeloma (MM) and sickle cell disease (SCD) using the Observational Medical Outcomes Partnership (OMOP) Common Data Model (CDM) and Fast Healthcare Interoperability Resources (FHIR) standards. The Data Hub also provides a framework for development of disease-specific learning communities (LC) and testing of health care delivery strategies. The ASH Research Collaborative SCD CTN is a clinical trials accelerator that creates efficiencies in the execution of multicenter clinical trials and has been initially developed for SCD. Both components are operational, with the Data Hub actively aggregating source data and the SCD CTN reviewing study candidates. This manuscript describes processes involved in developing core features of the ASH Research Collaborative to inform the stakeholder community in preparation for expansion to additional disease areas.

For avoiding the risk of information loss caused by quantum measurement and obtaining the concise and reliable fusion results of a given source weather dataset for decision-making, this article transforms a given source (weather) dataset into the collection of quantum states, and proposes a weather data detection and fusion method based on the von Neumann entropy controlled reduction of quantum representations. The reduction of the quantum representations within a data vector depends on the von Neumann entropy of this data vector. Then, all the source data samples (units) are classified into different subsets according to the gaps between the data samples recombined by the quantum elements after reduction. Depending on the impact factors of source data samples, the source data samples in a subset are fused into a new object data sample. After that, the reliabilities of all object data samples are evaluated through the predefined reliability evaluation model, and the object data samples with high reliabilities are useful for decision-making oriented applications (such as weather prediction). The experimental results demonstrate that the proposed data fusion method is more effective in obtaining the concise and reliable fusion results for decision-making oriented applications compared with the other mentioned data fusion methods.

Cross-domain scene classification identifies scene categories by learning knowledge from a labeled data set (source domain) to an unlabeled data set (target domain), where the source data and the target data are sampled from different distributions. A lot of domain adaptation methods are used to reduce the distribution shift across domains, and most existing methods assume that the source domain shares the same categories with the target domain. It is usually hard to find a source domain that covers all categories in the target domain. Some works exploit multiple incomplete source domains to cover the target domain. However, in such setting, the categories of each source domain are a subset of the target-domain categories, and the target domain contains ``unknown'' categories for each source domain. The existence of unknown categories results in the conventional domain adaptation unsuitable. Known and unknown categories should be treated separately. Therefore, a separation mechanism is proposed to separate the known and unknown categories in this article. First, multiple-source classifiers trained on the multiple source domains are used to coarsely separate the known/unknown categories in the target domain. The target images with high similarities to source images are selected as known categories, and the target images with low similarities are selected as unknown categories. Then, a binary classifier trained using the selected images is used to finely separate all target-domain images. Finally, only the known categories are implemented in the cross-domain alignment and classification. The target images get labels by integrating the hypotheses of multiple-source classifiers on the known categories. Experiments are conducted on three cross-domain data sets to demonstrate the effectiveness of the proposed method.

Summary Linguistic historiography analyzes how linguistic knowledge has been acquired, stored, used and diffused. This article examines what can happen if linguists rely on copies of source data rather than the source data itself. It takes as a case study linguistic data from Siraya, a now-extinct Formosan language. Documents compiled in the seventeenth century by Dutch missionaries in Taiwan form a significant source of data for Siraya. One such document, a wordlist known as the Utrecht Manuscript (UM), is the principal source for the lexicon of one variety of Siraya, “Siraya Proper”. It has been published three times. Each edition, however, contains many errors. These editions, rather than the manuscript, have been used by scholars investigating Siraya. This article aims to correct errors in the editions and secondary literature on the UM with my readings of the manuscript itself. It therefore presents a more accurate record of the lexicon of “Siraya Proper” as well as illustrating the importance of using primary rather than secondary sources of linguistic data. Finally, it introduces an online edition of the UM, which will provide scholars and language revivalists with a useful resource for this lexicon.

Objective Reducing suicidal behavior among patients in the healthcare system requires accurate and explainable predictive models of suicide risk across diverse healthcare settings. Materials and Methods We proposed a general targeted fusion learning framework that can be used to build a tailored risk prediction model for any specific healthcare setting, drawing on information fusion from a separate more comprehensive dataset with indirect sample linkage through patient similarities. As a proof of concept, we predicted suicide-related hospitalizations for pediatric patients in a limited statewide Hospital Inpatient Discharge Dataset (HIDD) fused with a more comprehensive medical All-Payer Claims Database (APCD) from Connecticut. Results We built a suicide risk prediction model for the source data (APCD) and calculated patient risk scores. Patient similarity scores between patients in the source and target (HIDD) datasets using their demographic characteristics and diagnosis codes were assessed. A fused risk score was generated for each patient in the target dataset using our proposed targeted fusion framework. With this model, the averaged sensitivities at 90% and 95% specificity improved by 67% and 171%, and the positive predictive values for the combined fusion model improved 64% and 135% compared to the conventional model. Discussion and Conclusions We proposed a general targeted fusion learning framework that can be used to build a tailored predictive model for any specific healthcare setting. Results from this study suggest we can improve the performance of predictive models in specific target settings without complete integration of the raw records from external data sources.

A prototype electromagnetic vibrator, referred to here as E-Vib, was upgraded and developed for broadband hardrock and mineral exploration seismic surveys. We selected the iron oxide mine in Blötberget, central Sweden, for a test site in 2019 for the newly developed E-Vib because of the availability of earlier seismic datasets (from 2015 to 2016) for verification of its performance for hardrock imaging purposes. The two-dimensional data acquisition consisted of a fixed geometry with 550 receiver locations spaced at every 5 m, employing both cabled and wireless seismic recorders, along an approximately 2.7 km long profile. The E-Vib operated at every second receiver station (i.e. 10 m spacing) with a linear sweep of 2–180 Hz and with a peak force of 7 kN. The processing workflow took advantage of the broadband signal generated by the E-Vib in this challenging hardrock environment with varying ground conditions. The processed seismic section shows a set of reflections associated with the known iron oxide mineralization and a major crosscutting reflection interpreted to be from a fault system likely to be crosscutting the mineralization. The broadband source data acquisition and subsequent processing helped to improve signal quality and resolution in comparison with the earlier workflows and data where a drophammer seismic source was used as the seismic source. These results suggest new possibilities for the E-Vib source for improved targeting in hardrock geological settings.

Supercomputer job management systems (JMS) are complex software which have a number of parameters and settings. Various simulating methods have been used in order to explore impact of such parameters on the JMS efficiency metrics. At the same time evaluating accuracy (adequacy) of the applied JMS models is one of the key points. The paper contains the results of adequacy measure experiments for various JMS models, including simulating with a virtual supercomputing nodes and with the Alea job scheduling simulator. JMS SUPPZ functioning at the Joint Supercomputer Center of the Russian Academy of Sciences (JSCC RAS) was used for the experiments. Source data for such simulating was created upon the statistics of supercomputer MVS–10P OP installed at JSCC RAS. The normalized Euclidean distance between the job residence (turnaround) time vectors, obtained from the job streams of the real supercomputer and JMS model, was used as a measure of adequacy. The experiments results have confirmed intuitive ideas about the studied simulating methods accuracy, that allows using the normalized Euclidean distance between the jobs turnaround times vectors as a measure of various JMS models adequacy.

Owing to constant progress of wireless communications, the Unmanned Aerial Vehicles (UAVs) routing networks (UAVs-RN) under mobile Internet of Things (MIoT) have been prevalent tools to deal with natural emergencies. But the achievement of effective responses and proper utility, still remains a challenging task. It is required to analyze multi-source data of UAVs-RN, so that optimal planning schemes under MIoT can be found. To bridge such gap, this work mainly takes three aspects of factors into consideration: rapid response, finite budget and uncertain signal fading. Accordingly, a data-driven intelligent planning model for UAVs-RN under MIoT, is put forward in this paper. Data about wildfire happened in local areas of Australia is selected to build experimental scenarios. And two kinds of UAVs, Surveillance and Situational Awareness drones and Radio Repeater drones, are considered in this study. Firstly, the source data is visualized and the internal trend is analyzed to verify true validity. Then, a multi-objective planning model is accordingly established to aggregate multi-source data. At last, a case study is deeply investigated on real-world data to assess the proposed approach and suggest feasible planning schemes.

One of the common problems in multiuser supercomputer systems is inaccurate user walltime request for a job. Usually, this time is significantly overestimated by the user. Scheduling algorithms implemented in modern supercomputer job management systems (JMS) are quite complicated and have a lot of settings. In this regard, the influence of the job walltime overestimate on the scheduling efficiency is not obvious. The job walltime prediction simulation can obtain the such influence estimation. The article contains simulation results when the predicting accuracy achieves 100 $$\%$$ . JMS referred to as SUPPZ is applied at the Joint Supercomputer Center of the RAS (JSCC RAS). SUPPZ was used for our simulations. Actual statistics of MVS–10P OP supercomputer installed at JSCC RAS was used for simulation as the source data. In this paper we used the SUPPZ simulation with virtual supercomputing nodes due to its highest accuracy. The simulation results showed a noticeable improvement of such metrics as average queue job wait time, average queue length and average slowdown.

In unsupervised domain adaptation (UDA), a classifier for a target domain is trained with labeled source data and unlabeled target data. Existing UDA methods assume that the source data come from the same source domain (i.e., single-source scenario) or from multiple source domains, whose feature spaces have the same dimension ( homogeneous ) but different distributions (i.e., multihomogeneous-source scenario). However, in the real world, for a specific target domain, we probably have multiple different-dimension ( heterogeneous ) source domains, which do not satisfy the assumption of existing UDA methods. To remove this assumption and move forward to a realistic UDA problem, this article presents a shared-fuzzy-equivalence-relation neural network (SFERNN) for addressing the multisource heterogeneous UDA problem. The SFERNN is a five-layer neural network containing $c$ source branches and one target branch. The network structure of the SFERNN is first confirmed by a novel fuzzy relation called multisource shared fuzzy equivalence relation. Then, we optimize parameters of the SFERNN via minimizing cross-entropy loss on $c$ source branches and the distributional discrepancy between each source branch and the target branch. Experiments distributed across eight real-world datasets are conducted to validate the SFERNN. This testing regime demonstrates that the SFERNN outperforms the existing single-source heterogeneous UDA methods, especially when the target domain contains few data.

Radar-based human activity recognition (HAR) finds various applications like assisted living and driver behavior monitoring. As radar data are heavily environment-dependent, it is becoming increasingly important to develop a transfer learning mechanism that enables a radar-based HAR system with desirable cross-environment adaptation feasibility. This paper concerns the issue of how radar-based HAR system can adapt to a new environment without source data. To this end, we devote to using the source hypothesis transfer learning architecture to build such an environment adaptation mechanism towards cross-environment radar-based HAR. In doing this, it is a challenging task to develop a reliable self-supervised labeling strategy for generating pseudo labels associated with the unlabeled target data, which is crucial to facilitate the learning of a target-specific feature extractor being responsible for environment adaptation. This paper presents the neighbor-aggregating-based labeling method and incorporates it with the existing clustering-based labeling method to perform the self-supervised labeling task. The logic behind our approach is that the above two labeling methods are complementary to each other in terms of making use of both local and global structures of adaptation data to supervise the labeling task. The coordination of both labeling methods is motivated to be implemented in the weighted combination form, which contributes to improving the reliability of generated labels. Experimental results on a public HAR dataset based on the frequency modulated continuous wave (FMCW) radar demonstrate the effectiveness of our approach.

Domain adaptation has attracted attention by leveraging knowledge from well-labeled source data to facilitate unlabeled target learning tasks. Numerous research efforts have been devoted to extracting effective features by incorporating the pseudolabels of target data. However, the transferable knowledge reflected by intradomain structure, interdomain correlation and label supervision has scarcely been considered simultaneously. In this paper, we propose a novel structure preservation adversarial network with target reweighting (SPTR) for unsupervised domain adaptation, in which local structure consistencies and category-level semantic alignment are simultaneously considered in the adversarial learning framework. Based on the labeled and pseudolabeled samples, we attempt to align both global and category-level domain statistics from different domains and simultaneously enforce structural consistency from feature space to label space in the source and target domains. Furthermore, to suppress the influence of falsely labeled target samples, a novel and generalized sample reweighting strategy is developed to assign target samples with different levels of confidence, which fully explores the knowledge of the target distribution to benefit the semantic transfer process. The experimental results in three transfer learning scenarios demonstrate the superiority of our proposed method over other state-of-the-art domain adaptation algorithms.

Widely acknowledged by experts, the inconsistency between the cells of the proton exchange membrane fuel cell stack during operation is an important cause of the fuel cell life decay. Existing studies mainly focus on qualitative analysis of the effects of operating parameters on fuel cell stack consistency. However, there is currently almost no quantitative research on predicting the voltage consistency through operating parameters with machine learning methods. To solve this problem, a three-dimensional model of proton exchange membrane fuel cell stack with five single cells is established in this paper. The Computational Fluid Dynamic (CFD) method is used to provide the source data for prediction model. After predicting the voltage consistency with several machine learning methods and comparing the accuracy through simulation data, the integrated regression method based on Gradient Boosting Decision Tree (GBDT) gets the highest score (0.896) and is proposed for quickly predicting the consistency of cell voltage through operating parameters. After verifying the GBDT method with the experimental data from the fuel cell stack of SUNRISE POWER, in which the accuracy score is 0.910, the universality and accuracy of the method is confirmed. The influencing sensitivity of each operating parameter is evaluated and the current density has the greatest influence on the predicted value, which accounts for 0.40. The prediction of voltage consistency under different combination of operating parameters can guide the optimization of structural parameters in the process of the fuel cell design and operating parameters in the process of fuel cell control.

Traditional machine learning methods assume that training and testing data must be from the same machine running condition (MRC) and drawn from the same distribution. However, in several real-time industrial applications, this assumption does not hold. The traditional methods work satisfactorily in steady-state conditions but fail in time-varying conditions. In order to utilize time-varying data in variable MRCs, this article proposes a novel low-level knowledge transfer framework using a deep neural network (DNN) model for condition monitoring of machines in variable running conditions. The low-level features have been extracted in time, frequency, and time–frequency domains. These features are extracted from the source data to train the DNN. The trained DNN-based parameters are then transferred to another DNN, which is modified according to the low-level features extracted from the target data. The proposed approach is validated through three case studies on: 1) the air compressor acoustic data set; 2) the Case Western Reserve University bearing data set; and 3) the intelligent maintenance system bearing data set. The prediction accuracy obtained for the above case studies is as high as 100%, 93.07%, and 100%, respectively, with fivefold cross-validation. These real-time results show considerable improvement in the prediction performance using the proposed approach. Note to Practitioners —Condition-based monitoring schemes are widely applicable to rotating machines in various industries since they operate in tough working situations, and consequently, unpredicted failures occur. These unpredicted failures may cause perilous accidents in the industries. CBM systems prevent such failures, which results in the reduction of equipment damage and, hence, increases machinery lifetime. Modern industries are so complex and generating huge data, and these data can be collected using sensors, but placing a large number of sensors is difficult and expensive for different but similar kinds of faults in industries. This also increases the cost due to additional sensors and circuits. In this article, the authors have proposed a novel low-level knowledge transfer framework using the deep neural network (DNN)-based method for condition monitoring of machines in variable running conditions. Low-level features have been extracted to reduce the computations of DNN drastically with improved performance. This article also considered additional faults in the target domain, which is more practical in real-time applications. The proposed scheme has been validated with three case studies on acoustic and vibration signatures.

Many superior person re-identification (re-ID) approaches face a challenge: their performance show disastrous when all supervised models are generalized to a new domain. Some researchers have recently proposed domain adaptive person re-ID methods to address this difficulty, whereas they directly leveraged the target-domain data that would generate lots of noisy pseudo labels. Hence, this paper proposes a distance constraint between features (DCF) method, which clusters the feature distribution fitted the real target-domain data. We assemble different parts of one person for multi-scale self-supervised learning. After introducing domain invariance and designing the inter-image and inter-class distance constraint to regulate distances between target samples, the feature distribution extracted from the encoder trained by the source data can fit the real target data distribution, which leads our domain adaptive model to enjoy the more reliable clustering results and thus obtain a great identification performance in target domain. Extensive experiments demonstrate our approach outperforms state-of-the-art methods on three large-scale released datasets.

The neutral model of stone procurement developed by Brantingham (2003, 2006) provides a formal means to investigate the formation of lithic discard patterning under changing forager mobility conditions. This study modifies Brantingham's (2006) Lévy walk model to examine the influence of discard probability on the spatial distribution of raw material abundance. The model outcome shows that forager movement and tool discard probability have similar effects on the simulated patterns of raw material transport, so it is difficult—if not impossible—to differentiate the respective influence of the two factors from distance to source distributions alone. This finding of equifinality complicates the task of interpretating hominin mobility from archaeological distance to source data, particularly in settings such as the Middle-Upper Paleolithic transition, which is marked by an important reorganization in hominin lithic technology that may have affected stone tool discard probability.

Problem Formulation in this research are indicates that How quality of service bidang Penempatan Tenaga Kerja on Dinas Tenaga Kerja Kota Banjarmasin consist of tangible, reliability, responsiveness, assurance and empathy up grade social satisfaction. Kind data concists of quantitative data and qualitative data. Source data consist of prymary data and secondary data. Draft this research according to method using survey research is collection of data only done for in part of population. Population and sample this research amount 150 respondents. Technique collected data among them Field research are questionnairy, interview, observation, documentation and library research. Technique data analysis be explained in descriptive with quantitative is statistics method to used for describe of data collected.

Acquisition technology of multisource data has outstanding advantages in enhancing collection efficiency and reducing cost. However, traditional seismic data processes cannot be applied to multisource blended data. Therefore, deblending technology of multisource data is the key to the research. In this letter, we propose a deblending method of multisource seismic data based on a periodically varying cosine code (PVCC). First, we design a PVCC to blend the seismic data. Next, the blending model is transformed into the minimum problem of the objective function. Then, the blended data are decomposed in the curvelet domain. Finally, the main source data are separated based on sparse inversion. Furthermore, we use the edge processing to eliminate the boundary effect in the processed seismic data. The examples of synthetic data and field data are adopted to demonstrate that the proposed method has great potential in the deblending of multisource data. In addition, the edge processing can effectively suppress the boundary effect.

The recent development in machine learning, especially in deep neural networks (DNN), has enabled learning-based end-to-end communication systems, where DNNs are employed to substitute all modules at the transmitter and receiver. In this article, two end-to-end frameworks for frequency-selective channels and multi-input and multi-output (MIMO) channels are developed, where the wireless channel effects are modeled with an untrainable stochastic convolutional layer. The end-to-end framework is trained with mini-batches of input data and channel samples. Instead of using pilot information to implicitly or explicitly estimate the unknown channel parameters as in current communication systems, the transmitter DNN learns to transform the input data in a way that is robust to various channel conditions. The receiver consists of two DNN modules used for channel information extraction and data recovery, respectively. A bilinear production operation is employed to combine the features extracted from the channel information extraction module and the received signals. The combined features are further utilized in the data recovery module to recover the transmitted data. Compared with the conventional communication systems, performance improvement has been shown for frequency-selective channels and MIMO channels. Furthermore, the end-to-end system can automatically leverage the correlation in the channels and in the source data to improve the overall performance.

Cross-project defect prediction (CPDP) refers to identifying defect-prone software modules in one project (target) using historical data collected from other projects (source), which can help developers find bugs and prioritize their testing efforts. Recently, CPDP has attracted great research interest. However, the source and target data usually exist redundancy and nonlinearity characteristics. Besides, most CPDP methods do not exploit source label information to uncover the underlying knowledge for label propagation. These factors usually lead to unsatisfactory CPDP performance. To address the above limitations, we propose a landmark selection-based kernelized discriminant subspace alignment (LSKDSA) approach for CPDP. LSKDSA not only reduces the discrepancy of the data distributions between the source and target projects, but also characterizes the complex data structures and increases the probability of linear separability of the data. Moreover, LSKDSA encodes label information of the source data into domain adaptation learning process and makes itself with good discriminant ability. Extensive experiments on 13 public projects from three benchmark datasets demonstrate that LSKDSA performs better than a range of competing CPDP methods. The improvement is $ 3.44\%-11.23\%$ in g-measure , $ 5.75\%-11.76\%$ in AUC , and $ 9.34\%-33.63\%$ in MCC , respectively.