Search

Most existing multiview clustering methods are based on the original feature space. However, the feature redundancy and noise in the original feature space limit their clustering performance. Aiming at addressing this problem, some multiview clustering methods learn the latent data representation linearly, while performance may decline if the relation between the latent data representation and the original data is nonlinear. The other methods which nonlinearly learn the latent data representation usually conduct the latent representation learning and clustering separately, resulting in that the latent data representation might be not well adapted to clustering. Furthermore, none of them model the intercluster relation and intracluster correlation of data points, which limits the quality of the learned latent data representation and therefore influences the clustering performance. To solve these problems, this article proposes a novel multiview clustering method via proximity learning in latent representation space, named multiview latent proximity learning (MLPL). For one thing, MLPL learns the latent data representation in a nonlinear manner which takes the intercluster relation and intracluster correlation into consideration simultaneously. For another, through conducting the latent representation learning and consensus proximity learning simultaneously, MLPL learns a consensus proximity matrix with k connected components to output the clustering result directly. Extensive experiments are conducted on seven real-world datasets to demonstrate the effectiveness and superiority of the MLPL method compared with the state-of-the-art multiview clustering methods.

Passenger-flow anomaly detection and prediction are essential tasks for intelligent operation of the metro system. Accurate passenger-flow representation is the foundation of them. However, spatiotemporal dependencies, complex dynamic changes, and anomalies of passenger-flow data bring great challenges to data representation. Taking advantage of the time-varying characteristics of data, we propose a novel passenger-flow representation model based on low-rank dynamic mode decomposition (DMD), which also integrates the global low-rank nature and sparsity to explore the spatiotemporal consistency of data and depict abrupt data, respectively. The model can detect anomalies and predict short-term passenger flow conveniently and flexibly. For anomaly detection, we further introduce a strong temporal Toeplitz regularization to characterize the temporal periodic change of data, so as to more accurately detect anomalies. We conduct experiments with smart card transaction data from the Beijing metro system to assess the performance of the model in two use cases. In terms of anomaly detection, the experimental results demonstrate that our method can detect anomalies efficiently, especially for time sequence anomalies. As for short-term prediction, our model is superior to other methods in most cases.

With the emergence of various techniques involved in deep learning the researchers of computer vision tends to focus on the strategies such as object recognition and segmentation of image. This has inclined to accomplish the deep learning techniques in 3D reconstruction of both specific and generic objects. As the space for reconstruction of 3D images either in single or multi view has envisioned the researchers to concentrate on the available technologies used for reconstruction. With the available built in methods and technologies in deep learning, the performance of the proposed methods were reviewed and analyzed using several parameters. As the remaking of 2D images is still in the beginning stage, it is important to study the 3D shape representations, various network architecture, methodologies and approaches behind 3D reconstruction. In this work a review of deep learning methods for single or multiple RGB images of specific and generic object 3D reconstruction was done. Several methods and their importance were also discussed along with the challenges encountered and with further research directions. This paper critically analyzes the various 3D Shaped Representations, 3D Data Network Architectures, Depth Estimation methods, Multi View Representations and the Data Representation Techniques.

Predicting treatment medicines is a key aspect of many intelligent healthcare systems. It's a very challenging task due to the following reasons: (1) heterogeneous nature of EHR data that typically include laboratory results, treatment medicines, disease conditions, and demographic details collected from disparate sources; (2) complex correlations among medical sequences, including inter-correlations between sequences and temporal intra-correlations within each sequence; (3) temporal diversity of these correlations, which is highly affected by changing disease progression. We proposes a dual adaptive sequential network, entitled DASNet, to dynamically predict treatment medicines for patients. Specifically, DASNet comprises the following three components. Decomposed Adaptive Long Short-Term Memory network (DA-LSTM) is designed to capture the intra- and inter-correlations in multiple heterogeneous temporal sequences. Then, we develop an Attentive Meta learning Network (AT-MetaNet), which produces location- and context-specific dynamic weight parameters for DA-LSTM, thus enabling it to model the time-varying multi-level correlations. Finally, we employ an ATtentive Fusion Network (AT-FuNet) to retrieve historical information and collectively fuse heterogeneous data representation embeddings to predict treatment medicines. The results of extensive experiments on the public MIMIC-III dataset covering 11 medical conditions demonstrate that the proposed end-to-end model can achieve the state-of-the-art prediction performance while providing clinically useful insights.

On-demand ride services, e.g., taxi companies, ridesourcing/transportation network companies (TNCs), play a vital role in nowadays transportation modes. The relocation/repositioning of idle vehicles is an important operational problem for service providers. With the recent success of deep learning, a large and growing body of literature emerged on learning-based transportation problems, including the relocation problem. In this work, we study the idle vehicle relocation problem to achieve better drivers' profits and customers' satisfaction. In detail, we formulate the learning problem on real-world road networks while recent repositioning works under a deep reinforcement learning scheme formulate the environment as grids. Meanwhile, a graph formulation enables learning with Graph Neural Network (GNN), which shows promising results in recent transportation studies. We empirically show that preserving the road network structure and learning the data representation with GNN contribute to improving the repositioning decision policy. Our simulation experiments are conducted on two real-world datasets, in New York City and Austin, respectively. The results show our formulation's advantages in terms of reductions of order rejection rates and passenger waiting time, and increasing of vehicle occupancy rates and driver revenues.

As residential photovoltaic (PV) system installations continue to increase rapidly, utilities need to identify the locations of these new components to manage the unconventional two-way power flow and maintain sustainable management of distribution grids. But historical records are unreliable and constant re-assessment of active residential PV locations is resource intensive. To resolve these issues, we propose to model the solar detection problem in a machine learning set up based on labeled data, e.g., supervised learning. However, the challenge for most utilities is limited labels or labels on only one type of users. Therefore, we design new semi-supervised learning and one-class classification methods based on autoencoders, which greatly improve the nonlinear data representation of human behavior and solar behavior. The proposed methods have been tested and validated not only on synthetic data based on a publicly available dataset, but also on real-world data from utility partners. The numerical results show robust detection accuracy, laying down the foundation for managing distributed energy resources in distribution grids.

Most previous EEG-based emotion recognition methods studied hand-crafted EEG features extracted from different electrodes. In this paper, we study the relation among different EEG electrodes and propose a deep learning method to automatically extract the spatial features that characterize the functional relation between EEG signals at different electrodes. Our proposed deep model is called ATtention-based LSTM with Domain Discriminator (ATDD-LSTM) that can characterize nonlinear relations among EEG signals of different electrodes. To achieve state-of-the-art emotion recognition performance, the architecture of ATDD-LSTM has two distinguishing characteristics: (1) By applying the attention mechanism to the feature vectors produced by LSTM, ATDD-LSTM automatically selects suitable EEG channels for emotion recognition, which makes the learned model concentrate on the emotion related channels in response to a given emotion; (2) To minimize the significant feature distribution shift between different sessions and/or subjects, ATDD-LSTM uses a domain discriminator to modify the data representation space and generate domain-invariant features. We evaluate the proposed ATDD-LSTM model on three public EEG emotional databases (DEAP, SEED and CMEED) for emotion recognition. The experimental results demonstrate that our ATDD-LSTM model achieves superior performance on subject-dependent (for the same subject), subject-independent (for different subjects) and cross-session (for the same subject) evaluation.

The development of Internet of Things systems (IoTs) and 5G technology has allowed image and text information to be collected and spread at an unprecedentedly high speed. To improve the data processing capabilities of IoTs, the semantic relations between images and text should be extracted efficiently and accurately. Therefore, to reduce the enormous semantic differences between images and text, existing methods introduce consensus knowledge graphs into image-text matching tasks. However, these methods result in noisy edges during the graph construction stage and overlook detailed knowledge extraction, leading to reduced performance in semantic matching. In this paper, a two-layer heterogeneous knowledge graph network is proposed to solve the above problems. The proposed model incorporates category knowledge and local knowledge for improved data representation. Specifically, a category-based hierarchical knowledge graph is constructed to learn representations of knowledge concepts through a hierarchical correlation graph embedding (HCGE) module. Then, a globally guided local attention (GLA) module is used to extract fine-grained local knowledge. Finally, the similarity between the input image and text is calculated based on knowledge-fused features to complete the matching process. Extensive experiments show that the proposed model can learn more effective knowledge features to improve the efficacy of image-text matching in.

Systems integrators and vendors of industrial components need to establish a security-by-design approach, which includes the assessment and subsequent treatment of security risks. However, conducting security risk assessments along the engineering process is a costly and labor-intensive endeavor due to the complexity of the system(s) under consideration and the lack of automated methods. This, in turn, hampers the ability of security analysts to assess risks pertaining to cyber-physical systems (CPSs) in an efficient manner. In this work, we propose a method that automatically identifies security risks based on the CPS's data representation, which exists within engineering artifacts. To lay the foundation for our method, we present security-focused semantics for the engineering data exchange format AutomationML (AML). These semantics enable the reuse of security-relevant know-how in AML artifacts by means of a formal knowledge representation, modeled with a security-enriched ontology. Our method is capable of automating the identification of security risk sources and potential consequences in order to construct cyber-physical attack graphs that capture the paths adversaries may take. We demonstrate the benefits of the proposed method through a case study and an open-source prototypical implementation. Finally, we prove that our solution is scalable by conducting a rigorous performance evaluation.

The plethora of ways of data representation and their applications to system modeling is inherently associated with dimensionality reduction. In a nutshell, the result of dimensionality reduction should support efficient ways of constructing ensuing models (classifiers, predictors) as well as an interpretation of the data themselves. Furthermore, there should be a suitable measure quantifying the quality of data positioned in the reduced space. We advocate that what makes the reduced data interpretable, goes hand in hand with revealing a logic fabric of the data, suppressing redundancy, and finally arriving at a logic description of data. The anticipation is that the reduced data can be described in the form of logic expressions formed over the original highly dimensional data. Evidently, having these above stated points in mind, the aim of this study is two-fold: (i) to develop a logic-oriented data representation, and (ii) to quantify the quality of results of dimensionality reduction by incorporating a facet of information granularity. In other words, we argue that the result of dimensionality reduction gives rise to information granules whose level of granularity associates with the quality of processing completed by the autoencoder. In light of the recent surge of architectures of deep learning, the study is focused on the construction and analysis of logic-oriented autoencoders. We propose a two-level architecture composed of the logic-oriented processing units (and processing carried out at the first layer of the autoencoder) followed by the or processing completed at the second layer. As data representation provided by the autoencoder is not ideal, we augment the original architecture by granular parameters which give rise to granular logic-oriented autoencoders. A suite of experiments is also reported.

In the SIKE implementation, the Montgomery representation has been mostly adopted in the finite field arithmetic computing as the corresponding reduction algorithm is considered the fastest method for implementing the modular reduction. In this paper, we propose a new data representation for the supersingular isogeny-based elliptic-curve cryptography (ECC), of which the SIKE is a subclass. The new representation can facilitate faster modular reduction implementation than the Montgomery reduction. Meanwhile, the other finite field arithmetic operations in the ECC can also benefit from the proposed data representation. We have implemented all the arithmetic operations in C language with constant execution time based on our proposed data representation and applied them to the newest SIKE software library. Targeting at the SIKEp751, we run our design and the optimized generic implementation on a 2.6GHz Intel Xeon E5-2690 processor. The experiment results show that for the parameters of SIKEp751, the proposed modular reduction algorithm is about 2.61x faster than the best Montgomery one and our scheme also performs significantly better for the other finite field operations. With these improvements, the overall software implementation for the SIKEp751 achieves about 1.65x speedup compared to the state-of-the-art generic implementatio

Information retrieval is a difficult process due to the overabundance of information on the web. Nowadays, search result responds to user queries with too many results although only a few are relevant. Therefore, the existing clustering methods that fail in clustering snippets (short texts) of web documents due to the low frequencies of document terms should be deeply investigated. One of the approaches that can be used to solve this problem is the expansion of document terms with semantically similar terms. Hence, a list of terms with their closest and accurate semantically similar words (word representation) must be built. This study aims to design and develop a new framework to enhance the performance of web search result clustering (WSRC). The research also presents a new unsupervised distributed word representation scheme where each word is represented by a vector of its semantically related words; such as scheme expands snippets and user queries. The proposed framework consists of several activities, such as (1) various standard datasets (Open Directory Project [ODP]-239 and MORESQUE) that are used for evaluating search result clustering algorithms for most cited dataset works, (2) text pre-processing, (3) document representation based on a new wiki-based k-nearest neighbors (KNN) representation method, (4) effect of the proposed model on the performance of traditional clustering methods (k-means, k-medoids, single-linkage, and complete-linkage) for WSRC, and (5) evaluation stage of the proposed method. Results indicate that enhanced clustering methods, according to the new wiki-KNN based representation method in comparison with the baseline methods, show a significant improvement in WSRC. Furthermore, the new data representation scheme has enhanced the overall performance of clustering methods.

Semi-supervised nonnegative matrix factorization (NMF) has received considerable attention in machine learning and data mining. A new semi-supervised NMF method, called dual semi-supervised convex nonnegative matrix factorization (DCNMF), is proposed in this paper for fully using the limited label information. Specifically, DCNMF simultaneously incorporates the pointwise and pairwise constraints of labeled samples as dual supervisory information into convex NMF, which results in a better low-dimensional data representation. Moreover, DCNMF imposes the nonnegative constraint only on the coefficient matrix but not on the base matrix. Consequently, DCNMF can process mixed-sign data, and hence enlarge the range of applications. We derive an efficient alternating iterative algorithm for DCNMF to solve the optimization, and analyze the proposed DCNMF method in terms of the convergence and computational complexity. We also discuss the relationships between DCNMF and several typical NMF based methods. Experimental results illustrate that DCNMF outperforms the related state-of-the-art NMF methods on nonnegative and mixed-sign datasets for clustering applications. This work is supported in part by the National Nature Science Foundation of China (No. U1701266, 91648208, 61976175), and Guangdong Intellectual Property Big Data Key Laboratory (No. 2018B030322016).

Recently, the convolutional neural network (CNN) has made great progress in hyperspectral image (HSI) classification because of its powerful feature extraction capability. However, the standard CNN based on grid sampling neglects the inherent relation between HSI data, which leads to poor regional edge delineation and generalization ability. Graph convolutional network (GCN) has been successfully applied to data representation in a non-Euclidean space, and it can extract discriminative embedded features by dynamically updating irregular graphs. In this letter, we propose a novel method termed attention mechanism-based dual-path convolutional network (AMDPCN), which is composed of a GCN-based global information learning model (GILM) and a CNN-based local feature extraction network (LFEN). Specifically, AMDPCN fuses the global spatial relationships explored by GILM and the local discriminant features extracted by LFEN with three different strategies: addition, multiplication, and concatenation. Furthermore, a multi-scale attention mechanism (MS-AM) is developed to mitigate the Hughes phenomenon by adaptive recalibrating the nonlinear interdependence among the features. Experiments on Kennedy Space Center and Indian Pines data sets demonstrate the advantages of the proposed AMDPCN to state-of-the-art methods.

Non-negative Tucker decomposition (NTD) is one of the most popular techniques for tensor data representation. To enhance the representation ability of NTD by multiple intrinsic cues, that is, manifold structure and supervisory information, in this article, we propose a generalized graph regularized NTD (GNTD) framework for tensor data representation. We first develop the unsupervised GNTD (UGNTD) method by constructing the nearest neighbor graph to maintain the intrinsic manifold structure of tensor data. Then, when limited must-link and cannot-link constraints are given, unlike most existing semisupervised learning methods that only use the pregiven supervisory information, we propagate the constraints through the entire dataset and then build a semisupervised graph weight matrix by which we can formulate the semisupervised GNTD (SGNTD). Moreover, we develop a fast and efficient alternating proximal gradient-based algorithm to solve the optimization problem and show its convergence and correctness. The experimental results on unsupervised and semisupervised clustering tasks using four image datasets demonstrate the effectiveness and high efficiency of the proposed methods.

Formal concept analysis, rough sets, and three-way decisions are prominent theories and methods for data representation and analysis. They have been applied to data mining, machine learning, artificial intelligence as well as many other areas. This special issue contains thirty high quality state-of-art research that addresses theoretic and applicational aspects of formal concept analysis, rough sets, and three-way decisions.

Building a visual overview of temporal event sequences with an optimal level-of-detail (i.e. simplified but informative) is an ongoing challenge - expecting the user to zoom into every important aspect of the overview can lead to missing insights. We propose a technique to build a multilevel overview of event sequences, whose granularity can be transformed across sequence clusters (vertical level-of-detail) or longitudinally (horizontal level-of-detail), using hierarchical aggregation and a novel cluster data representation Align-Score-Simplify. By default, the overview shows an optimal number of sequence clusters obtained through the average silhouette width metric - then users are able to explore alternative optimal sequence clusterings. The vertical level-of-detail of the overview changes along with the number of clusters, whilst the horizontal level-of-detail refers to the level of summarization applied to each cluster representation. The proposed technique has been implemented into a visualization system called Sequence Cluster Explorer (Sequen-C) that allows multilevel and detail-on-demand exploration through three coordinated views, and the inspection of data attributes at cluster, unique sequence, and individual sequence level. We present two case studies using real-world datasets in the healthcare domain: CUREd and MIMIC-III; which demonstrate how the technique can aid users to obtain a summary of common and deviating pathways, and explore data attributes for selected patterns., Comment: This is the author's version of the article to be published in IEEE Transactions on Visualization and Computer Graphics

In this paper, we design novel interactive deep learning methods to improve semantic interactions in visual analytics applications. The ability of semantic interaction to infer analysts' precise intents during sensemaking is dependent on the quality of the underlying data representation. We propose the $\text{DeepSI}_{\text{finetune}}$ framework that integrates deep learning into the human-in-the-loop interactive sensemaking pipeline, with two important properties. First, deep learning extracts meaningful representations from raw data, which improves semantic interaction inference. Second, semantic interactions are exploited to fine-tune the deep learning representations, which then further improves semantic interaction inference. This feedback loop between human interaction and deep learning enables efficient learning of user- and task-specific representations. To evaluate the advantage of embedding the deep learning within the semantic interaction loop, we compare $\text{DeepSI}_{\text{finetune}}$ against a state-of-the-art but more basic use of deep learning as only a feature extractor pre-processed outside of the interactive loop. Results of two complementary studies, a human-centered qualitative case study and an algorithm-centered simulation-based quantitative experiment, show that $\text{DeepSI}_{\text{finetune}}$ more accurately captures users' complex mental models with fewer interactions.

Recent years have witnessed the rapid development of face image hallucination techniques. However, the previous face hallucination methods are unsupervised and ignore the label information of training samples, leading to undesirable results. This article proposes a locality-constrained and category embedding representation (LCER) method to super-resolve face image in a supervised manner by embedding the label information in data representation. The proposed LCER incorporates the locality prior and category information into one unified framework, which aims to learn both the advantages of locality in preserving the true typologic structure of data manifold and the discriminability in exposing the class subspace information. Such strategy allows the LCER not only to preserve more sharpen image details but also to guarantee the face structure pattern be transferred mainly from the same subject in super-resolution reconstruction. Extensive experiments were conducted to evaluate the proposed LCER, and the comparative results demonstrate that it achieved superior face hallucination performance in both the quantitative measurements and visual impressions compared to several state-of-the-art.

Top-down proteomics analyzes intact proteoforms with all of their post-translational modifications and genetic and RNA splice variants. In addition, modifications introduced either deliberately or inadvertently during sample preparation, that is, via oxidation, alkylation, or labeling reagents, or through the formation of noncovalent adducts (e.g., detergents) further increase the sample complexity. To facilitate the recognition of protein modifications introduced during top-down analysis, we developed MSTopDiff, a software tool with a graphical user interface written in Python, which allows one to detect protein modifications by calculating and visualizing mass differences in top-down data without the prerequisite of a database search. We demonstrate the successful application of MSTopDiff for the detection of artifacts originating from oxidation, formylation, overlabeling during isobaric labeling, and adduct formation with cations or sodium dodecyl sulfate. MSTopDiff offers several modes of data representation using deconvoluted MS1 or MS2 spectra. In addition to artificial modifications, the tool enables the visualization of biological modifications such as phosphorylation and acetylation. MSTopDiff provides an overview of the artificial and biological modifications in top-down proteomics samples, which makes it a valuable tool in quality control of standard workflows and for parameter evaluation during method development.

Under the current artificial intelligence boom, machine translation is a research direction of natural language processing, which has important scientific research value and practical value. In practical applications, the variability of language, the limited capability of representing semantic information, and the scarcity of parallel corpus resources all constrain machine translation towards practicality and popularization. In this paper, we conduct deep mining of source language text data to express complex, high-level, and abstract semantic information using an appropriate text data representation model; then, for machine translation tasks with a large amount of parallel corpus, I use the capability of annotated datasets to build a more effective migration learning-based end-to-end neural network machine translation model on a supervised algorithm; then, for machine translation tasks with parallel corpus data resource-poor language machine translation tasks, migration learning techniques are used to prevent the overfitting problem of neural networks during training and to improve the generalization ability of end-to-end neural network machine translation models under low-resource conditions. Finally, for language translation tasks where the parallel corpus is extremely scarce but monolingual corpus is sufficient, the research focuses on unsupervised machine translation techniques, which will be a future research trend.

For landslide surface monitoring, the Global Navigation Satellite System (GNSS) has been widely used in landslides due to its real-time, all-weather, high-precision, simple operation and a high degree of automation. However, these data are not intuitive and visual data will be more interesting for users without professional knowledge. At the same time, the conventional data representation method is in the form of curves or tables for three-dimensional data of landslide surface deformation collected by GNSS. To make the data more intuitive, clear and valuable, it is easier for people to understand the process of landslide deformation and finally realize the visualization of decision. Here we show that a polar coordinate system rather than a Cartesian coordinate system is adopted to visualize the horizontal data, which not only shows the horizontal deformation of the landslide, but also easily knows the direction of the landslide deformation. The vertical data is in the form of slices rather than curves, which not only shows the deformation of the landslide surface, but also shows the process of the vertical change of the landslide in terms of the time series. Single GNSS monitoring station is composed of a GNSS receiver, GNSS antenna with random, a solar power unit, and a network transmission module. The system can be powered by the solar energy system, which can realize 24-hour unmanned operation, 7 days a week. The system can receive satellite signals in real-time process and analyze deformation data, then it automatically broadcast early warning information. Our results demonstrate that it is a better choice that the thematic map of Geographic Information System (GIS) is a technical system for collecting, storing, managing, calculating, analyzing and displaying geographic data supported by computer hardware and software systems. Here we show that the multi-dimensional properties of deformation monitoring and multiple expressions of the attribute values are displayed synchronously in order to obtain more useful information from the visual graphics.

Cross-domain data has become very popular recently since various viewpoints and different sensors tend to facilitate better data representation. In this article, we propose a novel cross-domain object representation algorithm (RLRCA) which not only explores the complexity of multiple relationships of variables by canonical correlation analysis (CCA) but also uses a low rank model to decrease the effect of noisy data. To the best of our knowledge, this is the first try to smoothly integrate CCA and a low-rank model to uncover correlated components across different domains and to suppress the effect of noisy or corrupted data. In order to improve the flexibility of the algorithm to address various cross-domain object representation problems, two instantiation methods of RLRCA are proposed from feature and sample space, respectively. In this way, a better cross-domain object representation can be achieved through effectively learning the intrinsic CCA features and taking full advantage of cross-domain object alignment information while pursuing low rank representations. Extensive experimental results on CMU PIE, Office-Caltech, Pascal VOC 2007, and NUS-WIDE-Object datasets, demonstrate that our designed models have superior performance over several state-of-the-art cross-domain low rank methods in image clustering and classification tasks with various corruption levels.

Low-precision data representation is important to reduce storage size and memory access for convolutional neural networks (CNNs). Yet, existing methods have two major limitations: (1) requiring re-training to maintain accuracy for deep CNNs and (2) needing 16-bit floating-point or 8-bit fixed-point for a good accuracy. In this article, we propose a low-precision (8-bit) floating-point (LPFP) quantization method for FPGA-based acceleration to overcome the above limitations. Without any re-training, LPFP finds an optimal 8-bit data representation with negligible top-1/top-5 accuracy loss (within 0.5%/0.3% in our experiments, respectively, and significantly better than existing methods for deep CNNs). Furthermore, we implement one 8-bit LPFP multiplication by one 4-bit multiply-adder and one 3-bit adder, and therefore implement four 8-bit LPFP multiplications using one DSP48E1 of Xilinx Kintex-7 family or DSP48E2 of Xilinx Ultrascale/Ultrascale+ family, whereas one DSP can implement only two 8-bit fixed-point multiplications. Experiments on six typical CNNs for inference show that on average, we improve throughput by over existing FPGA accelerators. Particularly for VGG16 and YOLO, compared to six recent FPGA accelerators, we improve average throughput by 3.5 and 27.5 and average throughput per DSP by 4.1 and 5 , respectively.

Shape classification from binary images is a challenging task within the computer vision community. Commonly, contour and structural features are computed to describe the objects and code patterns robust against rotation, scaling, and shape deformation. However, current techniques get a high-dimensional feature space decreasing the system performance and the attribute interpretability. Here, we introduce an enhanced and interpretable feature representation approach to support shape classification from binary images. Our method, named EIFR, employs a bag of contour fragments-based feature estimation, intrinsically robust to occlusion and shape deformation. Then, a ReliefF-based feature selection is applied to filter non-discriminative attributes. In turn, a kernel-alignment-based projection is used to measure the feature relevance enhancing the data representation through the matching between a similarity matrix computed from filtered attributes and a kernel matrix built from the shape labels. Attained results on benchmark datasets prove that EIFR improves the curvature-based features’ interpretability and favors the classification performance.

Abstract Background European cities are placing a larger emphasis on urban data consolidation and analysis for optimizing public transport in response to changing urban mobility dynamics. Despite the existing efforts, traffic data analysis often disregards vital situational context, including large-scale events, weather factors, traffic generation poles, social distancing norms, or traffic interdictions. Some of these sources of context data are still private, dispersed, or unavailable for the purpose of planning or managing urban mobility. Addressing the above observation, the Lisbon city Council has already established efforts for gathering historic and prospective sources of situational context in standardized semi-structured repositories, triggering new opportunities for context-aware traffic data analysis. Research questions The work presented in this paper aims at tackling the following main research question: How to incorporate historical and prospective sources of situational context into descriptive and predictive models of urban traffic data? Methodology We propose a methodology anchored in data science methods to integrate situational context in the descriptive and predictive models of traffic data, with a focus on the three following major spatiotemporal traffic data structures: i) georeferenced time series data; ii) origin-destination tensor data; iii) raw traffic event data. Second, we introduce additional principles for the online consolidation and labelling of heterogeneous sources of situational context from public repositories. Third, we quantify the impact produced by situational context aspects on public passenger transport data gathered from smart card validations along the bus (CARRIS), subway (METRO) and bike sharing (GIRA) modes in the city of Lisbon. Results The gathered results stress the importance of incorporating historical and prospective context data for a guided description and prediction of urban mobility dynamics, irrespective of the underlying data representation. Overall, the research offers the following major contributions: A novel methodology on how to acquire, consolidate and incorporate different sources of context for the context-enriched analysis of traffic data; The instantiation of the proposed methodology in the city of Lisbon, discussing the role of recent initiatives for the ongoing monitoring of relevant context data sources within semi-structured repositories, and further showing how these initiatives can be extended for the context-sensitive modelling of traffic data for descriptive and predictive ends; A roadmap of practical illustrations quantifying impact of different context factors (including weather, traffic interdictions and public events) on different transportation modes using different spatiotemporal traffic data structures; and A review of state-of-the-art contributions on context-enriched traffic data analysis. The contributions reported in this work are anchored in the empirical observations gathered along the first stage of the ILU project (see footnote 1), providing a study case of interest to be followed by other European cities.

Advanced technologies are required in future mobile wireless networks to support services with highly diverse requirements in terms of high data rate and reliability, low latency, and massive access. Deep Learning (DL), one of the most exciting developments in machine learning and big data, has recently shown great potential in the study of wireless communications. In this article, we provide a literature review on the applications of DL in the physical layer. First, we analyze the limitations of existing signal processing techniques in terms of model accuracy, global optimality, and computational scalability. Next, we provide a brief review of classical DL frameworks. Subsequently, we discuss recent DL-based physical layer technologies, including both DL-based signal processing modules and end-to-end systems. Deep neural networks are used to replace a single or several conventional functional modules, whereas the objective of the latter is to replace the entire transceiver structure. Lastly, we discuss the open issues and research directions of the DL-based physical layer in terms of model complexity, data quality, data representation, and algorithm reliability.

The viral transmission in public places and transportations can be minimized by following the world health organization (WHO) guideline. However, the uncertainty in a dynamic system complicates the social engagement to the physical distancing regulation. This study aims to overcome this obstacle in MRT stations and train by developing an adaptive queue line system. The system was developed using low-cost hardware and open-source software to guide passengers using visual information. The system works by capturing seat images and identify the presence of humans using a cloud machine learning service. The physical representation of MRT was translated to data representation using the internet of things (IoT). The data then streamed using an asynchronous API with a representative endpoint. The endpoint is then accessed by a display computer in the destination station platform to provide visual information. The visual information was ergonomically designed with visual display principles, including the minimum content load, layout, color combination, and dimension of contents. The design of the system was evaluated by Markov simulation of virus transmission in train and usability testing of the visual design. The implementation of the system has balanced the queue line capacity in station and crowd spots distribution in MRT. The system was effective due to the visual cortex manipulation by visual information. Consequently, the aerosol and falling droplets' viral transmission radius can be reduced. Accordingly, the chance for airborne transmission can be lowered. Therefore, the adaptive queue line system is a non-pharmaceutical intervention of viral transmission diseases in public transportation

In this work, we present a novel bitsliced high-performance Viterbi algorithm suitable for high-throughput and data-intensive communication. A new column-major data representation scheme coupled with the bitsliced architecture is employed in our proposed Viterbi decoder that enables the maximum utilization of the parallel processing units in modern parallel accelerators. With the help of the proposed alteration of the data scheme, instead of the conventional bit-by-bit operations, 32-bit chunks of data are processed by each processing unit. This means that a single bitsliced parallel Viterbi decoder is capable of decoding 32 different chunks of data simultaneously. Here, the Viterbi’s Add-Compare-Select procedure is implemented with our proposed bitslicing technique, where it is shown that the bitsliced operations for the Viterbi internal functionalities are efficient in terms of their performance and complexity. We have achieved this level of high parallelism while keeping an acceptable bit error rate performance for our proposed methodology. Our suggested hard and soft-decision Viterbi decoder implementations on GPU platforms outperform the fastest previously proposed works by and , achieving 21.41 and 8.24 Gbps on Tesla V100, respectively.

In this work we address the problem of 3D object detection from point clouds in data-limited environments. Training with simulated data is a common approach in such scenarios; however a sim-to-real gap exists between clean and crisp simulated clouds and noisy real clouds. Previous sim-to-real approaches for processing point cloud scenes have compressed clouds into 2D and used 2D transfer techniques. However, this may compress useful 3D information and does not effectively reason about the unstructured nature of point cloud data. We thus propose a 3D adversarial training architecture that leverages an adaptive sampling module to reason about the unstructured nature of point cloud data. Our approach encourages the 3D feature encoder to extract features that are invariant across simulated and real scenes. We validate our approach in the context of the DARPA Subterranean Challenge and demonstrate that our 3D adversarial training architecture improves 3D object detection performance by up to 15% depending on the data representation.

There are over 1 million transgender people living in the United States, and 33% report negative experiences with a healthcare provider, many of which are connected to data representation in electronic health records (EHRs). We present recommendations and common pitfalls involving sex- and gender-related data collection in EHRs. Our recommendations leverage the needs of patients, medical providers, and researchers to optimize both individual patient experiences and the efficacy and reproducibility of EHR population-based studies. We also briefly discuss adequate additions to the EHR considering name and pronoun usage. We add the disclaimer that these questions are more complex than commonly assumed. We conclude that collaborations between local transgender and gender-diverse persons and medical providers as well as open inclusion of transgender and gender-diverse individuals on terminology and standards boards is crucial to shifting the paradigm in transgender and gender-diverse health.

This book provides a structured introduction of the key concepts and techniques that enable in-/near-memory computing. For decades, processing-in-memory or near-memory computing has been attracting growing interest due to its potential to break the memory wall. Near-memory computing moves compute logic near the memory, and thereby reduces data movement. Recent work has also shown that certain memories can morph themselves into compute units by exploiting the physical properties of the memory cells, enabling in-situ computing in the memory array. While in- and near-memory computing can circumvent overheads related to data movement, it comes at the cost of restricted flexibility of data representation and computation, design challenges of compute capable memories, and difficulty in system and software integration. Therefore, wide deployment of in-/near-memory computing cannot be accomplished without techniques that enable efficient mapping of data-intensive applications to such devices, without sacrificing accuracy or increasing hardware costs excessively. This book describes various memory substrates amenable to in- and near-memory computing, architectural approaches for designing efficient and reliable computing devices, and opportunities for in-/near-memory acceleration of different classes of applications.

Concept factorization (CF) is an effective matrix factorization model which has been widely used in many applications. In CF, the linear combination of data points serves as the dictionary based on which CF can be performed in both the original feature space as well as the reproducible kernel Hilbert space (RKHS). The conventional CF treats each dimension of the feature vector equally during the data reconstruction process, which might violate the common sense that different features have different discriminative abilities and therefore contribute differently in pattern recognition. In this paper, we introduce an auto-weighting variable into the conventional CF objective function to adaptively learn the corresponding contributions of different features and propose a new model termed Auto-Weighted Concept Factorization (AWCF). In AWCF, on one hand, the feature importance can be quantitatively measured by the auto-weighting variable in which the features with better discriminative abilities are assigned larger weights; on the other hand, we can obtain more efficient data representation to depict its semantic information. The detailed optimization procedure to AWCF objective function is derived whose complexity and convergence are also analyzed. Experiments are conducted on both synthetic and representative benchmark data sets and the clustering results demonstrate the effectiveness of AWCF in comparison with some related models.

During the past years, multi-view clustering algorithms have demonstrated satisfactory clustering results by fusing the multiple views of the dataset. Nowadays, the researches of dimensionality reduction and learning discriminative features from multi-view data have soared in the literatures. As for clustering, generating the suitable subspace of the high dimensional multi-view data is crucial to boost the clustering performance. In addition, the relationship between the original data and the clusters still remains uncovered. In this article, we design a new multi-view low rank sparse representation method based on three-way clustering to tackle these challenges, which derive the common consensus low dimensional representation from the multi-view data and further proceed to get the relationship between the data items and clusters. Specifically, we accomplish this goal by taking advantage of the low-rank and the sparse factor on the data representation matrix. The $$L_{2,1}$$ norm is imposed on error matrix to reduce the impact of noise contained in the data. Finally, a new objective function is constructed to preserve the consistency between the views by using the low-rank sparse representation technique. The weighted low-rank matrix is utilized to build the consensus low rank matrix. Then, the whole objective function is optimized by using the Augmented Lagrange’s Multiplier algorithm. Further, to find the uncertain relationship between the data items and the clusters, we pursue the neighborhood based three-way clustering technique to reflect the data items into core and fringe regions. Experiments conducted on the real-world datasets show the superior performance of the proposed method compared with the state-of-the-art algorithms.

The availability of Electronic Health Records (EHR) in health care settings has provided tremendous opportunities for early disease detection. While many supervised learning models have been adopted for EHR-based disease early detection, the ill-posed inverse problem in the parameter learning has imposed a significant challenge on improving the accuracy of these algorithms. In this paper, we propose CRLEDD – Causality-Regularized Learning for Early Detection of Disease , an algorithm to improve the performance of Linear Discriminant Analysis (LDA) on top of diagnosis-frequency vector data representation. While most existing regularization methods exploit sparsity regularization to improve detection performance, CRLEDD provides a unique perspective by ensuring positive semi-definiteness of the sparsified precision matrix used in LDA which is different from the regular regularization method (e.g., L2 regularization). To achieve this goal, CRLEDD employs Graphical Lasso to estimate the precision matrix in the ill-posed settings for enhanced accuracy of LDA classifiers. We perform extensive evaluation of CRLEDD using a large-scale real-world EHR dataset to predict mental health disorders (e.g., depression and anxiety) of college students from 10 universities in the U.S. We compare CRLEDD with other regularized LDA and downstream classifiers. The result shows that CRLEDD outperforms all baselines in terms of accuracy and F1 scores.

Clustering is a crucial but challenging task in data mining and machine learning. Recently, deep clustering, which derives inspiration primarily from deep learning approaches, has achieved state-of-the-art performance in various applications and attracted considerable attention. Nevertheless, most of these approaches fail to effectively learn informative cluster-oriented features for data with spatial correlation structure, e.g., images. To tackle this problem, in this paper, we develop a deep convolutional self-paced clustering (DCSPC) method. Specifically, in the pretraining stage, we propose to utilize a convolutional autoencoder to extract a high-quality data representation that contains the spatial correlation information. Then, in the finetuning stage, a clustering loss is directly imposed on the learned features to jointly perform feature refinement and cluster assignment. We retain the decoder to avoid the feature space being distorted by the clustering loss. To stabilize the training process of the whole network, we further introduce a self-paced learning mechanism and select the most confident samples in each iteration. Through comprehensive experiments on seven popular image datasets, we demonstrate that the proposed algorithm can consistently outperform state-of-the-art rivals.

Data availability all datasets used in this study are available on the web at the following addresses (see also the manuscript):- https://www.usgs.gov/volcanoes/kilauea/monitoring- https://www.usgs.gov/observatories/hawaiian-volcano-observatory- https://pnsn.org/seismograms- https://www.pnsn.org/tremor- http://www.fdsn.org/networks/detail/Y4_2014.; International audience; Data visualization, and to a lesser extent data sonification, are classic tools to the scientific community. However, these two approaches are very rarely combined, although they are highly complementary: our visual system is good at recognizing spatial patterns, whereas our auditory system is better tuned for temporal patterns. In this article, data representation methods are proposed that combine visualization, sonification, and spatial audio techniques, in order to optimize the user’s perception of spatial and temporal patterns in a single display, to increase the feeling of immersion, and to take advantage of multimodal integration mechanisms. Three seismic data sets are used to illustrate the methods, covering different physical phenomena, time scales, spatial distributions, and spatio-temporal dynamics. The methods are adapted to the specificities of each data set, and to the amount of information that the designer wants to display. This leads to further developments, namely the use of audification with two time scales, the switch from pure audification to time-modulated noise, and the switch from pure audification to sonic icons. First user feedback from live demonstrations indicates that the methods presented in this article seem to enhance the perception of spatio-temporal patterns, which is a key parameter to the understanding of seismically active systems, and a step towards apprehending the processes that drive this activity.

Non-negative Tucker decomposition (NTD) has been developed as a crucial method for non-negative tensor data representation. However, NTD is essentially an unsupervised method and cannot take advantage of label information. In this paper, we claim that the low-dimensional representation extracted by NTD can be treated as the predicted soft-clustering coefficient matrix and can therefore be learned jointly with label propagation in a unified framework. The proposed method can extract the physically-meaningful and parts-based representation of tensor data in their natural form while fully exploring the potential ability of the given labels with a nearest neighbors graph. In addition, an efficient accelerated proximal gradient (APG) algorithm is developed to solve the optimization problem. Finally, the experimental results on five benchmark image data sets for semi-supervised clustering and classification tasks demonstrate the superiority of this method over state-of-the-art methods.

The decision-making process is always associated with uncertainty and vagueness because of the incomplete decision data and unpredictable decision-making behavior. The concurrence of the epistemic uncertainty and the aleatory uncertainty, especially when only the qualitative data are accessible, poses more challenges to the decision-making. The probabilistic linguistic term set provides a flexible tool for data representation in such an environment. However, most of the existing decision-making methods for the probabilistic linguistic term sets depend on the aggregation operator. The subjectivity of the decision-makers and the decision-making behaviors are not fully considered. To solve this problem, a quantum-like probabilistic linguistic term decision-making framework is proposed, in which the quantum state, quantum data processing, and quantum interference effect are analyzed in detail. The decision-making beliefs of decision-makers are in a superposition state and their subjective relationships are depicted in the interference term. A multi-attribute decision-making method in this framework is then proposed for practical applications. Then the applications for sea-battlefield situation assessment are presented as the illustration and validation. A comparative analysis is also provided to depict the advantages of the proposed method, based on which appropriate conclusions and remarks are drawn.

Purpose In the literature on automated phenotyping of chronic obstructive pulmonary disease (COPD), there is a multitude of isolated classical machine learning and deep learning techniques, mostly investigating individual phenotypes, with small study cohorts and heterogeneous meta-parameters, e.g., different scan protocols or segmented regions. The objective is to compare the impact of different experimental setups, i.e., varying meta-parameters related to image formation and data representation, with the impact of the learning technique for subtyping automation for a variety of phenotypes. The identified associations of these parameters with automation performance and their interactions might be a first step towards a determination of optimal meta-parameters, i.e., a meta-strategy. Methods A clinical cohort of 981 patients (53.8 ± 15.1 years, 554 male) was examined. The inspiratory CT images were analyzed to automate the diagnosis of 13 COPD phenotypes given by two radiologists. A benchmark feature set that integrates many quantitative criteria was extracted from the lung and trained a variety of learning algorithms on the first 654 patients (two thirds) and the respective algorithm retrospectively assessed the remaining 327 patients (one third). The automation performance was evaluated by the area under the receiver operating characteristic curve (AUC). 1717 experiments were conducted with varying meta-parameters such as reconstruction kernel, segmented regions and input dimensionality, i.e., number of extracted features. The association of the meta-parameters with the automation performance was analyzed by multivariable general linear model decomposition of the automation performance in the contributions of meta-parameters and the learning technique. Results The automation performance varied strongly for varying meta-parameters. For emphysema-predominant phenotypes, an AUC of 93%-95% could be achieved for the best meta-configuration. The airways-predominant phenotypes led to a lower performance of 65%-85%, while smooth kernel configurations on average were unexpectedly superior to those with sharp kernels. The performance impact of meta-parameters, even that of often neglected ones like the missing-data imputation, was in general larger than that of the learning technique. Advanced learning techniques like 3D deep learning or automated machine learning yielded inferior automation performance for non-optimal meta-configurations in comparison to simple techniques with suitable meta-configurations. The best automation performance was achieved by a combination of modern learning techniques and a suitable meta-configuration. Conclusions Our results indicate that for COPD phenotype automation, study design parameters such as reconstruction kernel and the model input dimensionality should be adapted to the learning technique and may be more important than the technique itself. To achieve optimal automation and prediction results, the interaction between input those meta-parameters and the learning technique should be considered. This might be particularly relevant for the development of specific scan protocols for novel learning algorithms, and towards an understanding of good study design for automated phenotyping.

This paper describes a program—SPARQL Query Generator (SQG)—which takes as input an OWL ontology, a set of object descriptions in terms of this ontology and an OWL class as the context, and generates relatively large numbers of queries about various types of descriptions of objects expressed in RDF/OWL. The intent is to use SQG in evaluating data representation and retrieval systems from the perspective of OWL semantics coverage. While there are many benchmarks for assessing the efficiency of data retrieval systems, none of the existing solutions for SPARQL query generation focus on the coverage of the OWL semantics. Some are not scalable since manual work is needed for the generation process; some do not consider (or totally ignore) the OWL semantics in the ontology/instance data or rely on large numbers of real queries/datasets that are not readily available in our domain of interest. Our experimental results show that SQG performs reasonably well with generating large numbers of queries and guarantees a good coverage of OWL axioms included in the generated queries.

A novel sub-part learning scheme is introduced in our work for the purpose of recognizing handwritten numeral images. The idea is borrowed from the concept of visual perception and part-wise integration of visual information by the cortical regions of the brain. In this context, each numeral image is divided into four half-parts: top-half, bottom-half, left-half and right-half; the other half of the image being kept masked. An efficient data representation is derived in an unsupervised manner, from each image part, using convolutional auto-encoders (CAE), for our learning scheme that involves both early and late fusion of features. The chief advantage of the features derived from convolutional auto-encoders is the preservation of 2D spatial locality while the features are being filtered layer-by-layer through the convolutional architecture. The features derived from each individual CAE are fused by concatenation in our early fusion scheme, and learnt using an appropriate classifier. The late fusion strategy involves learning the probability density pertaining to the predicted values emanating from the four base classifiers using a meta-learner classifier. The early-cum-late fusion is proposed in the later stage of our work to combine the goodness of both schemes and enhance the performance. The support vector machine is used in all the classification stages. Experiments on the benchmark MNIST dataset of handwritten English numerals prove that our method competes favorably to the state of the art, as inferred from the high classification scores achieved. Our method thus provides a computationally simple and effective methodology for sub-part learning and part-wise integration of information from different parts of the image. The method also contributes to saving in computational expense since, at a time, only a small part of the image is processed, speeding up the inferencing process.

The Yanbu'a method is applied at the Darul Ulum Kudus Islamic Boarding School to improve learning outcomes to read al-Qur’an. This study aims to determine how to learn al-Qur’an using the Yanbu'a method at the Darul Ulum Kudos Islamic Boarding School. This type of research is descriptive and qualitative research. Data collection techniques were carried out through interviews and documents. The data analysis used followed the analysis of Miles and Huberman's model, namely data reduction, data representation and verification. The results of this study indicate that: First, on the basis of active, innovative, creative, effective and interesting learning the Yanbu'a method is adopted at the Darulum Suci Islamic Boarding School. Second, the role of Yanbu language learning at Darul Ulum Islamic Boarding School to study al-Qur’an is very important to improve learning outcomes to read al-Qur’an. Based on the research results, this method can improve students' ability to read al-Qur’an based on makharijul and tajwid letters.

The rapid growth of Internet of Things (IoT) and sensing technologies has led to an increasing interest in time-series data analysis. In many domains, detecting patterns of IoT data and interpreting these patterns are challenging issues. There are several methods in time-series analysis that deal with issues such as volume and velocity of IoT data streams. However, analysing the content of the data streams and extracting insights from dynamic IoT data is still a challenging task. In this paper, we propose a pattern representation method which represents time-series frames as vectors by first applying Piecewise Aggregate Approximation (PAA) and then applying Lagrangian Multipliers. This method allows representing continuous data as a series of patterns that can be used and processed by various higher-level methods. We introduce a new change point detection method which uses the constructed patterns in its analysis. We evaluate and compare our representation method with Blocks of Eigenvalues Algorithm (BEATS) and Symbolic Aggregate approXimation (SAX) methods to cluster various datasets. We have also evaluated our proposed change detection method. We have evaluated our algorithm using UCR time-series datasets and also a healthcare dataset. The evaluation results show significant improvements in analysing time-series data in our proposed method.

This article is devoted to the basic models of descriptive image analysis, which is the leading section of the modern mathematical theory of image analysis and recognition. The fundamental problem that the subject of the article addresses is automated extraction of information from images necessary for making intelligent decisions. Descriptive analysis envisages the implementation of image analysis processes in the image formalization space, the elements of which are various forms (states, phases) of the image representation, which is converted from the original form to one convenient for recognition (i.e., to a model), and data representation conversion models. The image analysis processes are considered to be sequences of transformations implemented in the phase space and providing the construction of phase states of the image, which together form the phase trajectory of image transfer from the original form to the model. The study of the image formalization space leads to formalization of the concepts of representation/image model, as well as to construction of models of the image recognition and analysis processes and formulation of mathematical statements of image recognition and analysis problems. Two types of image analysis models are considered: (1) models reflecting the methodology and mathematical foundations of image recognition and analysis—the problem statement, used mathematical and heuristic methods, algorithmic content of the process: (a) a model based on reverse algebraic closure, (b) a model based on the property of equivalence of images of the same class, (c) a model based on multiple image models and multiple classifiers; (2) models characterizing the architecture and structure of the recognition process: (a) a multilevel model for combining algorithms and initial data for image recognition, (b) information structure for generating descriptive algorithmic image recognition schemes. A brief description of the above models is given. A comparative analysis of the relationships and specifics of the models is carried out. Directions for further research are discussed.

Massive volumes of high-dimensional data that evolve over time are continuously collected by contemporary information processing systems, which bring up the problem of organizing these data into clusters, i.e. achieving the purpose of dimensional reduction, and meanwhile learning their temporal evolution patterns. In this paper, a framework for evolutionary subspace clustering, referred to as LSTM–ESCM, is introduced, which aims at clustering a set of evolving high-dimensional data points that lie in a union of low-dimensional evolving subspaces. In order to obtain the parsimonious data representation at each time step, we propose to exploit the so-called self-expressive trait of the data at each time point. At the same time, LSTM networks are implemented to extract the inherited temporal patterns behind data in the overall time frame. An efficient algorithm has been proposed. Numerous experiments are carried out on real-world datasets to demonstrate the effectiveness of our proposed approach. The results show that the suggested algorithm dramatically outperforms other known similar approaches in terms of both run time and accuracy.

Semantic interpretation of multi-modal datasets is of great importance in many domains of geospatial data analysis. However, when training models for automated semantic segmentation, labeled training data is required and in case of multi-modality for each representation form of the scene. To completely avoid the time-consuming and cost-intensive involvement of an expert in the annotation procedure, we propose an Active Learning (AL) pipeline where a Random Forest classifier selects a subset of points sufficient for training and where necessary labels are received from the crowd. In this AL loop, we aim on coupled semantic segmentation of an Airborne Laser Scanning (ALS) point cloud and the corresponding 3D textured mesh generated from LiDAR data and imagery in a hybrid manner. Within this work we pursue two main objectives: i) We evaluate the performance of the AL pipeline applied to an ultra-high resolution ALS point cloud and a derived textured mesh (both benchmark datasets are available at https://ifpwww.ifp.uni-stuttgart.de/benchmark/hessigheim/default.aspx). ii) We investigate the capabilities of the crowd regarding interpretation of 3D geodata and observed that the crowd performs about 3 percentage points better when labeling meshes compared to point clouds. We additionally demonstrate that labels received solely by the crowd can power a machine learning system only differing in Overall Accuracy by less than 2 percentage points for the point cloud and less than 3 percentage points for the mesh, compared to using the completely labeled training pool. For deriving this sparse training set, we ask the crowd to label 0.25 % of available training points, resulting in costs of 190 $.

As the core machining process element in the overall manufacturing process of a part, the machining process route plays an important role in improving final manufacturing quality. In the current CAPP system, the decision-making of the process route still depends on human-computer interaction and essentially depends on human intelligence. In the past decade, deep learning technology architecture has been gradually improved, which provides a new enabling technology for intelligent process planning. Recently, some researchers have applied deep learning to process route decision-making. However, due to the challenges of data representation and deep learning network construction, this promising solution is still at infancy. To address the two challenges, this paper presents a novel process route generation approach based on deep learning. First, we propose a fourth-order tensor model to represent the geometry and technological requirements of a part. And the relation matrix is constructed to represent the relationships among machining features. The process route is represented as a sequential set of one-hot vectors. Then, we construct an encoder-decoder neural architecture to automatically generate the machining process route for the part. The 3D convolution neuron network-based encoder converts the geometry, technological requirements, and the information of the relationships among machining features into a higher layer of vector representation, and the long short-term memory network-based decoder maps this representation to the process route. The whole neural architecture including the encoder and decoder is jointly trained to maximize the conditional probability of the target process route given the training part. Finally, the paper takes slot cavity parts as examples to verify the feasibility and effectiveness of the proposed approach.

As the first attempt, this article considers how to provide a design methodology for style clustering on stylistic data, where each cluster depends on both the similarities between data samples and its latently or apparently distinguishable style. By taking our previous fuzzy k plane clustering algorithm as the basic framework, a fuzzy style k-plane clustering (S-KPC) algorithm is proposed to have its distinctive merits: First, the nuances between styles of clusters can be well identified by using the proposed twofold data representation. That is to say, style matrices are used to express the structure, hence style information of each cluster, whereas the augmentation of the original features of data with enhanced nodes is taken as an abstract representation so as to move the manifold structure of data apart. Such a twofold data representation can make us realize S-KPC readily in an incremental way. Second, by means of alternating optimization strategy, the objective function of S-KPC can be optimized such that each discriminant function of each cluster shares the advantages of both simple regression models and functional-link neural networks. Extensive experiments on synthetic and real-world datasets demonstrate that S-KPC has comparable clustering performance with several compared methods on the adopted ordinary datasets, and yet it obviously outperforms them on stylistic datasets.

Cyber-physical-social big data generated from ubiquitous devices and diverse spaces generally are multi-source, heterogeneous, and deeply intertwined. To efficiently analyze and handle the ubiquitous cyber-physical-social big data, tensor is considered as an effective tool, but the curse of dimensionality is still the main bottleneck of tensor-based big data analysis. Tensor networks can considerably alleviate or overcome it through the tensor approximate theory. Therefore, this paper focuses on developing an efficient big data processing framework based on tensor networks and providing an incremental tensor train decomposition approach for the streaming big data. Concretely, this paper first presents a hierarchical cyber-physical-social big data processing framework composed of three planes, namely, data representation and decomposition, data storage and processing, and data analysis and service, in which tensor train (TT) and quantized TT decompositions are particularly introduced to remarkably overcome the curse of dimensionality. Besides, to efficiently handle the continuous streaming big data and avoid the repeated decomposition for the history data, an incremental tensor train decomposition (ITTD) approach is proposed and the complexities are further analyzed in detail. Experimental results demonstrate that ITTD demonstrably outperforms the nonincremental TT decomposition in execution time on the precise of guaranteeing the nearly equal approximation error.