Search

Tumor detection supported by Raman spectroscopy is becoming increasingly popular, yet the relevance of spectral variation and feature selection retains unclear. Here we determined the correlation and difference between spectral characteristic and feature evaluation for leukocytes and tumor cells. Some peaks were found to show noticeable spectral differences, and their intensity distributions were investigated, finding using log-normal distribution to describe Raman intensity pattern may be more appropriate. Further the importance of all Raman features was calculated, where some other peak features occupied the top status. By surveying the intensity variation and feature evaluation for those peaks, we concluded the peak with the highest importance does not correspond to the peak location with the most noticeable intensity difference in spectra. Moreover, the peak intensity ratio of I1517/I719 associated with protein to nucleic acid level presented the maximum separation, thus, it can be recognized as a special indicator to develop an alternative cancer detection. It is inspiring to introduce advanced statistical models into bio-spectroscopic fields but those intrinsic spectral variations rather than classification performance should be valued. Our explorations can provide possibilities to reveal the essences within tumor carcinogenesis based on Raman spectroscopy, further overwhelming the obstacles during the translation into clinical applications.

The thermal efficiency limitation of the Flat-plate Micro Heat Pipe (FMHP) is a major challenge in the development of the FMHP, where the effect of wick structure and wettability on its thermal performance is studied to improve the thermal efficiency of the FMHP. In this work, a copper spiral woven mesh and copper foam Composite Wick FMHP (CW-FMHP) is designed based on the conventional Copper Foam Wick FMHP (CFW-FMHP), and its thermal performance is analyzed regarding the wick structure and internal gas–liquid two-phase flow characteristics. An oxidized copper spiral woven mesh and copper foam Composite Wick FMHP (OCW-FMHP) has been further developed through the modification of composite wick wettability. The performance tests are carried out with the thermal transfer characteristics of CW-FMHP, OCW-FMHP, and CFW-FMHP under different filling rates and different thermal powers. The experimental results show that the thermal transfer performance of CW-FMHP reaches the optimal under a liquid filling rate of 150%, where the maximum thermal power is 15.7 W, 35.3% higher than that of the CFW-FMHP under the same filling rate. Moreover, the dynamic response characteristics of the CW-FMHP are significantly improved. The thermal resistance of the CW-FMHP is 0.48 °C/W under the filling rate of 150% at the thermal power of 10 W with a reduction of 9.4% compared to the CFW-FMHP under the same condition. Furthermore, the optimal filling rate for OCW-FMHP is lower compared with the CW-FMHP. The maximum thermal power of OCW-FMHP increases to 17.8 W while the thermal resistance reduces to 0.34 °C/W under the liquid filling rate of 140%. This implies that the composite wick structure designed in this work can improve the thermal transfer performance of the FMHP, and the composite wick with wettability modification is more effective regarding both thermal resistance and maximum thermal power.

A concept of Texas Instrument (TI)-Phase-only Spatial Light Modulator (PLM) is reported which utilizing the dynamic piston motion of PLM pixels to form a discretized blazed grating. By fully manipulating the piston motion of pixels and increasing the available discretization level of the quasi-blazed grating, the Diffraction Efficiency (DE) can achieve close to 99%, which qualifies PLM an ideal candidate for beam steering functionality combining with its MEMS based high refresh rate and large aperture. The DE of the discretized blazed grating is proven to have 86% with 633 nm monochromatic light incident at 25° with 16 discretization levels and 2𝜋 round-trip phase modulation by RCWA algorithm. Furthermore, additional factors which lead to the degradation of diffraction efficiency is also analyzed.

The use of containers in microservices has gained popularity as it facilitates agile development, resource governance, and software maintenance. Container reallocation aims to achieve workload balance via reallocating containers over physical machines. It affects the overall performance of microservice-based systems. However, container scheduling and reallocation remain an open issue due to their complexity in real-world scenarios. In this paper, we propose a novel Multi-Phase Local Search (MPLS) algorithm to optimize container reallocation. The experimental results show that our optimization algorithm outperforms state-of-the-art methods. In practice, it has been successfully applied to Microsoft 365 system to mitigate hotspot machines and balance workloads across the entire system.

In cloud systems, an accurate capacity planning is very important for cloud provider to improve service availability. Traditional methods simply predicting "when the available resources is exhausted" are not effective due to customer demand fragmentation and platform allocation constraints. In this paper, we propose a novel prediction approach which proactively predicts the level of resource allocation failures from the perspective of low capacity status. By jointly considering the data from different sources in both time series form and static form, the proposed approach can make accurate LCS predictions in a complex and dynamic cloud environment, and thereby improve the service availability of cloud systems. The proposed approach is evaluated by real-world datasets collected from a large scale public cloud platform, and the results confirm its effectiveness.

There has been a rapidly increasing demand for developing highly configurable software systems, which urgently calls for effective testing methods. In practice, t-wise coverage has been widely recognized as a useful metric to evaluate the quality of a test suite for testing highly configurable software systems, and achieving high t-wise coverage is important for ensuring test adequacy. However, state-of-the-art methods usually cost a fairly long time to generate large test suites for high pairwise coverage (i.e., 2-wise coverage), which would lead to ineffective and inefficient testing of highly configurable software systems. In this paper, we propose a novel local search based sampling approach dubbed LS-Sampling for achieving high t-wise coverage. Extensive experiments on a large number of public benchmarks, which are collected from real-world, highly configurable software systems, show that LS-Sampling achieves higher 2-wise and 3-wise coverage than the current state of the art. LS-Sampling is effective, since on average it achieves the 2-wise coverage of 99.64% and the 3-wise coverage of 97.87% through generating a small test suite consisting of only 100 test cases (90% smaller than the test suites generated by its state-of-the-art competitors). Furthermore, LS-Sampling is efficient, since it only requires an average execution time of less than one minute to generate a test suite with high 2-wise and 3-wise coverage.

Capacity management has always been a great challenge for cloud platforms due to massive, heterogeneous on-demand instances running at different times. To better plan the capacity for the whole platform, a class of cloud computing instances have been released to collect computing demands beforehand. To use such instances, users are allowed to submit jobs to run for a pre-specified uninterrupted duration in a flexible range of time in the future with a discount compared to the normal on-demand instances. Proactively scheduling those pre-collected job requests considering the capacity status over the platform can greatly help balance the computing workloads along time. In this work, we formulate the scheduling problem for these pre-collected job requests under uncertain available capacity as a Prediction + Optimization problem with uncertainty in constraints, and propose an effective algorithm called Controlling under Uncertain Constraints (CUC), where the predicted capacity guides the optimization of job scheduling and job scheduling results are leveraged to improve the prediction of capacity through Bayesian optimization. The proposed formulation and solution are commonly applicable for proactively scheduling problems in cloud computing. Our extensive experiments on three public, industrial datasets shows that CUC has great potential for supporting high reliability in cloud platforms.

Decomposition approach is an important component in multi-objective evolutionary algorithm based on decomposition (MOEA/D), which is a popular method for handing many-objective optimization problems (MaOPs). This paper presents a theoretical analysis on the convergence ability of using the typical weighted sum (WS), Tchebycheff (TCH) or penalty-based boundary intersection (PBI) approach in a basic MOEA/D for solving two benchmark MaOPs. The results show that using WS, the algorithm can always find an optimal solution for any subproblem in polynomial expected runtime. In contrast, the algorithm needs at least exponential expected runtime for some subproblems if using TCH or PBI. Moreover, our analyses discover an obvious shortcoming of using WS, that is, the optimal solutions of different subproblems are easily corresponding to the same solution. In addition, this analysis indicates that if using PBI, a small value of the penalty parameter is a good choice for faster converging to the Pareto front, but it may lose the diversity. This study reveals some optimization behaviors of using three typical decomposition approaches in the well-known MOEA/D framework for solving MaOPs.

The oil-immersed transformer is a crucial piece of equipment in the power system. Operating at the specified temperature is necessary to ensure the normal operation of the transformer. The insulation paper on the winding surface has a significant impact on the actual temperature of the transformers, which is often overlooked by researchers. The one-dimensional steady-state heat conduction model of the transformer is established by analyzing the heat diffusion process of winding to transformer oil. Atomic force microscope was used to observe the microsurface structure of insulation paper and copper. According to the experiment, the heat transfer resistance in the series process of heat transfer at [Formula: see text]C is 0.0138 m2 K/W. Space thermal circuit model of transformer is established by thermoelectricity analogy method, and the simulation circuit is optimized according to the boundary conditions set up in the actual environment. The results show that the error of the hot spot temperature is closer to the measured temperature and decreases by 2.5% when considering the thermal resistance of insulation paper.

Click-through rate (CTR) prediction aims to recall the advertisements that users are interested in and to lead users to click, which is of critical importance for a variety of online advertising systems. In practice, CTR prediction is generally formulated as a conventional binary classification problem, where the clicked advertisements are positive samples and the others are negative samples. However, directly treating unclicked advertisements as negative samples would suffer from the severe label noise issue, since there exist many reasons why users are interested in a few advertisements but do not click. To address such serious issue, we propose a reinforcement learning based noise filtering approach, dubbed RLNF, which employs a noise filter to select effective negative samples. In RLNF, such selected, effective negative samples can be used to enhance the CTR prediction model, and meanwhile the effectiveness of the noise filter can be enhanced through reinforcement learning using the performance of CTR prediction model as reward. Actually, by alternating the enhancements of the noise filter and the CTR prediction model, the performance of both the noise filter and the CTR prediction model is improved. In our experiments, we equip 7 state-of-the-art CTR prediction models with RLNF. Extensive experiments on a public dataset and an industrial dataset present that RLNF significantly improves the performance of all these 7 CTR prediction models, which indicates both the effectiveness and the generality of RLNF.

Purpose The purpose of this study is to explore the effects of review group’s content-related and environment-associated attributes on information adoption intention of information readers. In addition, this study further investigates the effects of these determinants in different websites with different commercial attributes (i.e. online stores and third party forums). Design/methodology/approach This study used a structured online survey to collect data from an online store and a third party forum in China; totally 302 responses were collected. Findings The empirical results confirm that all of the five determinants significantly affect the information adoption intention of information readers. Furthermore, the authors found that four of the determinants have distinct effects in different websites. Originality/value The findings of this study validate the significant moderating role of website attributes in readers’ information processing. Information readers use distinct criteria to evaluate the received review information; electronic word-of-mouth determinants have varying effects on different websites with different commercial attributes.

Feature selection is a meaningful aspect of data mining that aims to select more relevant data features and provide more concise and explicit data descriptions. It is beneficial for constructing an effective learning model and reducing the consumption of memory and time. In real-life applications, imbalanced data are ubiquitous, such as those in medical diagnoses, intrusion detection, and credit ratings. In recent years, feature selection for imbalanced data has attracted increasing research attention. Neighborhood rough set theory has been effectively applied to feature selection when dealing with mixed types of data. In this study, we propose an approach for feature selection for imbalanced data employing neighborhood rough set theory. The significance of features is defined by carefully studying the upper and lower boundary regions. The uneven distribution of the classes is considered during the definition of the feature significance. A discernibility-matrix-based feature selection method, which is a key method in rough set theory, is used; then, a novel algorithm for feature selection (RSFSAID) is proposed. The uncertainty of feature selection resulting from different parameters is investigated, and a particle swarm optimization algorithm is used to determine the optimized parameters in the algorithm. Extensive experiments are performed with public datasets to evaluate the proposed method. Experimental results show that the RSFSAID algorithm can improve the classification performance of imbalanced data compared to four other algorithms.

Dominance-based Rough Set Approach (DRSA) is widely applied in processing the information and data with a preference-order relation. In many real-life Ordered Information Systems (OIS), the attribute values and objects may vary simultaneously with time. Existing incremental updating algorithms in OIS are all single-dimensional, i.e., they only consider the individual variation of attribute values or objects. In addition, all the objects need to be compared with each other during updating. In this paper, we focus on designing specific multi-dimensional algorithms to efficiently update approximations on the simultaneous variation of attribute values and objects in OIS. The properties of the P-generalized decision in DRSA are firstly presented. Then two novel notions e.g., the P-generalized decision lower and upper domains, are defined, respectively. These two notions indicate the practical dominance dependence between different objects in OIS, and provide a simplified definition of the P-generalized decision. The approach based on these two notions can greatly reduce the comparisons between different objects, as well as provide novel strategies which can efficiently obtain the updated P-generalized decision to obtain the updated approximations. Furthermore, we develop two domain-wise algorithms which correspond to two cases of the multi-dimensional variation of attribute values and objects in OIS, respectively. Moreover, in implementing the algorithms, the sorting strategies are integrated to solve the problem of time-consuming traversals. A series of experimental results illustrate that our proposed domain-wise algorithms are not only more efficient than both of the traditional static algorithm and the integrated single-dimensional algorithms for dealing with the multi-dimensional variation of attribute values and objects, but also more efficient than the single-dimensional algorithms for dealing with the individual variation of objects in OIS.

Accurate short-term traffic flow prediction plays an indispensable role for solving traffic congestion. However, the structure of traffic data is nonlinear and complicated. It is a challenge to get high precision. The least square support vector machine (LSSVM) has powerful capabilities for time series and nonlinear regression prediction problems if it can select appropriate parameters. To search the optimal parameters of LSSVM, this paper proposes a hybrid optimization algorithm which combines particle swarm optimization (PSO) with genetic algorithm. The main contributions are twofold: (1) A hybrid optimization method is proposed, which can skip the local optimal pitfall with less learning time by introducing a selection strategy, crossover and mutation operators into PSO; (2) the crossover and mutation operators are controlled by adaptive probability functions. The crossover and mutation probabilities increase when the population fitness is concentrated, and decrease when the fitness is dispersed. It can effectively improve the precision and speed of convergence. The proposed model is verified based on the measured data. The experimental results show that our new model yields better prediction ability and relatively high computational efficiency compared with other related models.

Integrated optical waveguide sensor has characteristics of good insulation, broadband response, small size, and large dielectric strength, which is appropriate for measurement of intensive pulsed electric field. Two types of integrated optics sensors using through and reflective type asymmetric Mach–Zehnder interferometer are designed, fabricated, and implemented. Key parameters of the optical waveguide and antenna are theoretically analyzed and designed. The time domain response, input/output characteristics and frequency response of the sensors are investigated. The temperature stability is evaluated by measuring output power fluctuations of the sensors. Experiment results reveal that the sensitivity of the through type sensor is a little higher than that of the reflective type sensor, while the size (length) of the reflective type sensor is only half of the through type. Besides, as only a single optical fiber is coupled, it is convenient for the reflective type sensor to measure in very small space. Applications of the sensor for measurement of the lightning electromagnetic pulse, nuclear electromagnetic pulse, and pulsed radio frequency fields are demonstrated. The measurement results indicate that the developed integrated optics sensor has powerful potential and is expected to be widely used for measurement of intense pulsed electric fields.

Three-way decisions theory has been developed and applied to decision making in uncertain environments. Decision tasks under ambiguity involving both missing values and hierarchical concept are abundant in real-world applications. In this paper, we study the update problem of three-way decisions with dynamic variation of scales in incomplete multi-scale information systems. The updating mechanisms of decision granules induced by the similarity relation are exploited with the cut refinement and coarsening through the attribute value taxonomies, and the dynamic tendencies of conditional probability are hence estimated with evolving granularity structure. Reasonable updating behaviors of probabilistic parameters are characterized in the framework of decision-theoretic rough sets, in order to relate optimism and pessimism decision making on the basis of the variation in attitudes to losses or costs of decisions during the periodic transformation between coarser granular-scale and refiner granular-scale of data. Then the dynamic tendencies of conditional probability with the updating behaviors of probabilistic parameters together are incorporated into the incremental process of updating three-way decisions. Theoretical justifications and semantic interpretations are provided to guarantee the satisfaction of the proposed method for incremental three-way decisions in multi-scale incomplete information systems. Experimental results on several UCI datasets show that the proposed incremental algorithms efficiently update three-way decisions for handling transformation of data scales caused by the cut refinement and coarsening through the attribute value taxonomies.

Kai Xin San (KXS, containing ginseng, hoelen, polygala, and acorus), a traditional Chinese herbal compound, has been found to regulate cognitive dysfunction; however, its mechanism of action is still unclear. In this study, 72 specific-pathogen-free male Kunming mice aged 8 weeks were randomly divided into a vehicle control group, scopolamine group, low-dose KXS group, moderate-dose KXS group, high-dose KXS group, and positive control group. Except for the vehicle control group and scopolamine groups (which received physiological saline), the doses of KXS (0.7, 1.4 and 2.8 g/kg per day) and donepezil (3 mg/kg per day) were gastrointestinally administered once daily for 2 weeks. On day 8 after intragastric treatment, the behavioral tests were carried out. Scopolamine group and intervention groups received scopolamine 3 mg/kg per day through intraperitoneal injection. The effects of KXS on spatial learning and memory, pathological changes of brain tissue, expression of apoptosis factors, oxidative stress injury factors, synapse-associated protein, and cholinergic neurotransmitter were measured. The results confirmed the following. (1) KXS shortened the escape latency and increased residence time in the target quadrant and the number of platform crossings in the Morris water maze. (2) KXS increased the percentage of alternations between the labyrinth arms in the mice of KXS groups in the Y-maze. (3) Nissl and terminal deoxynucleotidyl transferase-mediated dUTP nick end labeling staining revealed that KXS promoted the production of Nissl bodies and inhibited the formation of apoptotic bodies. (4) Western blot assay showed that KXS up-regulated the expression of anti-apoptotic protein Bcl-2 and inhibited the expression of pro-apoptotic protein Bax. KXS up-regulated the expression of postsynaptic density 95, synaptophysin, and brain-derived neurotrophic factor in the cerebral cortex and hippocampus. (5) KXS increased the level and activity of choline acetyltransferase, acetylcholine, superoxide dismutase, and glutathione peroxidase, and reduced the level and activity of acetyl cholinesterase, reactive oxygen species, and malondialdehyde through acting on the cholinergic system and reducing oxidative stress damage. These results indicate that KXS plays a neuroprotective role and improves cognitive function through reducing apoptosis and oxidative stress, and regulating synapse-associated protein and cholinergic neurotransmitters.

Information fusion is capable of fusing and transforming multiple data derived from different sources to provide a unified representation for centralized knowledge mining that facilitates effective decision-making, classification and prediction, etc. Multisource interval-valued data, characterizing the uncertainty phenomena in the data in the form of intervals in different sources, are the most common symbolic data which widely exist in many real-world applications. This paper concentrates on efficient fusing of multisource interval-valued data with the dynamic updating of data sources, involving the addition of new sources and deletion of obsolete sources. We propose a novel data fusion method based on fuzzy information granulation, which translates multisource interval-valued data into trapezoidal fuzzy granules. Given this effectively fusing capability, we develop incremental mechanisms and algorithms for fusing multisource interval-valued data with a dynamic variation of data sources. Finally, extensive experiments are carried out to verify the effectiveness of the proposed algorithms when comparing to six different fusion algorithms. Experimental results show that the proposed fusion method outperforms other related approaches. Furthermore, the proposed incremental fusion algorithms can reduce the computing overhead in comparison with the static fusion algorithm when adding and deleting multiple data sources.

Networks are ubiquitous in the real world. Link prediction, as one of the key problems for network-structured data, aims to predict whether there exists a link between two nodes. The traditional approaches are based on the explicit similarity computation between the compact node representation by embedding each node into a low-dimensional space. In order to efficiently handle the intensive similarity computation in link prediction, the hashing technique has been successfully used to produce the node representation in the Hamming space. However, the hashing-based link prediction algorithms face accuracy loss from the randomized hashing techniques or inefficiency from the learning to hash techniques in the embedding process. Currently, the Graph Neural Network (GNN) framework has been widely applied to the graph-related tasks in an end-to-end manner, but it commonly requires substantial computational resources and memory costs due to massive parameter learning, which makes the GNN-based algorithms impractical without the help of a powerful workhorse. In this paper, we propose a simple and effective model called #GNN, which balances the trade-off between accuracy and efficiency. #GNN is able to efficiently acquire node representation in the Hamming space for link prediction by exploiting the randomized hashing technique to implement message passing and capture high-order proximity in the GNN framework. Furthermore, we characterize the discriminative power of #GNN in probability. The extensive experimental results demonstrate that the proposed #GNN algorithm achieves accuracy comparable to the learning-based algorithms and outperforms the randomized algorithm, while running significantly faster than the learning-based algorithms. Also, the proposed algorithm shows excellent scalability on a large-scale network with the limited resources.

In cloud computing, provisioning virtual machines (VMs) fast and reliably is a fundamental yet challenging problem, particularly in cloud environments of changing workloads and shifting demand patterns. A trivial solution of instantiating a VM from scratch per customer request may not satisfy business SLA's and would degrade customer experience. Hence, provisioning VMs in advance motivated with machine learning (ML) infused into the cloud computing system to predict upcoming VM request demands is an advisable solution. In this paper, we first describe a number of system integration challenges including 1) how to achieve low latency provisioning to quickly adjust to customer demand pattern shifts, 2) how to efficiently scale to serve a large number of VM configurations supported in the cloud environment, and 3) how to reliably consume recommended prediction results for VM provision despite of anticipated operation failures and timeout. We then present the high level solution design with discussions of our developed system to address the aforementioned challenges. Our system has been implemented and deployed successfully in Microsoft Azure exhibiting significant improvements for VM provisioning experience with regards to latency and reliability requirements.

Combinatorial interaction testing (CIT) is a popular approach to detecting faults in highly configurable software systems. The core task of CIT is to generate a small test suite called a t-way covering array (CA), where t is the covering strength. A major drawback of existing solvers for CA generation is that they usually need considerable time to obtain a high-quality solution, which hinders its wider applications. In this paper, we describe FastCA, an effective and efficient tool for generating constrained CAs. We observe that the high time consumption of existing meta-heuristic algorithms is mainly due to the procedure of score computation. To this end, we present a much more efficient method for score computation. Thanks to this new lightweight score computation method, FastCA can work in the gradient mode to effectively explore the search space. Experiments on a broad range of real-world benchmarks and synthetic benchmarks show that FastCA significantly outperforms state-of-the-art solvers, in terms of both the size of obtained covering array and the run time. Video: https://youtu.be/-6CuojQIt-k Repository: https://github.com/jkunlin/FastCATool.git

Combinatorial interaction testing (CIT) is an important technique for testing highly configurable software systems with demonstrated effectiveness in practice. The goal of CIT is to generate test cases covering the interactions of configuration options, under certain hard constraints. In this context, constrained covering arrays (CCAs) are frequently used as test cases in CIT. Constrained Covering Array Generation (CCAG) is an NP-hard combinatorial optimization problem, solving which requires an effective method for generating small CCAs. In particular, effectively solving t-way CCAG with t ≥ 4 is even more challenging. Inspired by the success of automated algorithm configuration and automated algorithm selection in solving combinatorial optimization problems, in this paper, we investigate the efficacy of automated algorithm configuration and automated algorithm selection for the CCAG problem, and propose a novel, automated CCAG approach called AutoCCAG. Extensive experiments on public benchmarks show that AutoCCAG can find much smaller-sized CCAs than current state-of-the-art approaches, indicating the effectiveness of AutoCCAG. More encouragingly, to our best knowledge, our paper reports the first results for CCAG with a high coverage strength (i.e., 5-way CCAG) on public benchmarks. Our results demonstrate that AutoCCAG can bring considerable benefits in testing highly configurable software systems.

With millions of customers accessing online service all over the world, ensuring high service availability is very critical for cloud system. In recent years, empowered by advanced data mining and machine learning technology, there emerges extensive study on data-driven solution to detect anomalous system behavior and diagnose the root cause. However, without any surveilance of data generation process, the existing passive data-driven approach may lead to biased analysis result induced by observed and unobserved confounding factors in the dynamic and heterogeneous system, and thus affect service availability with misleading mitigation actions. In this paper, we propose to infuse causal thinking to the current data-driven solution for cloud system. We developed CARE, a causal aware root cause discovery engine, which utilizes Random Control Trial to proactively generate less ambiguous data for further analysis. A case study shows the application of CARE to Microsoft Office365.

The stability and effective functioning of a power grid depends upon the reliability as well as safety prospects of a transformer which is the principal component of the power system. The natural oil circulation in the oil-immersed transformer drifts the oil flow by the change in the density which is again caused by the change in temperature. But the pressure is comparatively low with respect to the oil pump causing the lower down of velocity which in turn dissipates unimpressionable heat. In this paper, a noble attempt has been made to build a natural oil circulation layer-type winding transformer based on a fluid-solid coupling heat transfer model pertaining to the internal temperature field and flow field by using the fluid mechanics’ simulation software. In the next stage, the authors have analyzed the temperature distribution, hot spot, position condition of the variable winding, the horizontal the oil channel width as well as the temperature of the winding transformer. The paper is of great significance as it examines the influence of the flow characteristics of oil channel on the winding temperature distribution. The outcome can significantly improve the structure of heat dissipation and the operating environment of the transformer.

With the rapid deployment of cloud platforms, high service reliability is of critical importance. An industrial cloud platform contains a huge number of disks, and disk failure is a common cause of service unreliability. In recent years, many machine learning based disk failure prediction approaches have been proposed, and they can predict disk failures based on disk status data before the failures actually happen. In this way, proactive actions can be taken in advance to improve service reliability. However, existing approaches treat each disk individually and do not explore the influence of the neighboring disks. In this paper, we propose Neighborhood-Temporal Attention Model (NTAM), a novel deep learning based approach to disk failure prediction. When predicting whether or not a disk will fail in near future, NTAM is a novel approach that not only utilizes a disk’s own status data, but also considers its neighbors’ status data. Moreover, NTAM includes a novel attention-based temporal component to capture the temporal nature of the disk status data. Besides, we propose a data enhancement method, called Temporal Progressive Sampling (TPS), to handle the extreme data imbalance issue. We evaluate NTAM on a public dataset as well as two industrial datasets collected from millions of disks in Microsoft Azure. Our experimental results show that NTAM significantly outperforms state-of-the-art competitors. Also, our empirical evaluations indicate the effectiveness of the neighborhood-ware component and the temporal component underlying NTAM as well as the effectiveness of TPS. More encouragingly, we have successfully applied NTAM and TPS to Microsoft cloud platforms (including Microsoft Azure and Microsoft 365) and obtained benefits in industrial practice.

The recently reported “Angular Spatial Light Modulator” (ASLM) light engine, using pulsed illumination synchronized to a Digital Micromirror Device (DMD), shows significant promise to enhance pixel counts of Near-to-Eye Displays (NED) without increasing package volume, but requires an uncommon illumination driver. We present a field effect transistor based constant-current driver that is fast, compact, and scalable to RGB illumination. The digital-to-analog convertor modulates intensity on-the-fly for illumination-based multiplexing. The driver outputs 100 ns pulses, up to 24 kHz repetition rate. The circuit is demonstrated for two laser diodes and for two LEDs in an ASLM-enhanced pixel count display.

A holographic lidar concept established on the Digital Micromirror Mirror (DMD)-based hybrid light modulation is reported, which multiplexes coarse-steering by sawtooth phase modulation and fine-steering by binary amplitude modulation. The hybrid steering is achieved by overlaying displayed Computer-Generated Holograms (CGHs) with a sawtooth blazed grating phase mask, which the blaze angle programmed by synchronized short-pulse illumination of transitioning micromirrors creating the CGHs. The steering principle is demonstrated as a 2D beam steering scheme with a 532 nm visible pulse laser, and implemented into a 905 nm lidar system with a 44° field-of-view, 0.9°×0.4° angular resolution, 7.8 FPS video frame rate, and 1 m detection distance.

Attribute reduction is a significant process of data preprocessing to overcome the challenges posed by multidimensional data analysis. Missing values in the data are usually unavoidable in the real applications, so it is important to select features with high importance efficiently in incomplete data. The theory of rough sets has been widely used in attribute reduction for uncertain data mining. To enable the rough set theory for large-scale incomplete data analysis, this paper develops a novel distributed attribute reduction algorithm based on Apache Spark cluster computing platform. By taking the advantage of positive approximation technique for reducing the data broadcasting gradually while reducing each redundant attribute iteratively, the proposed algorithm can significantly accelerate the attribute reduction in leveraging a computer cluster when processing large-scale incomplete data. Numerical experiments on different UCI data sets evidences the proposed parallel algorithm achieves high performance in terms of extensibility and scalability.

In the domestic production of recycled plastics, it is difficult to separate olefin polymers such as polyethylene and polypropylene. Therefore, the prerequisite for identifying the cargo-solid waste properties of recycled plastics is to determine the polymer composition. In this paper, Fourier transform infrared spectroscopy was used to detect polypropylene content in polyethylene and polypropylene blends. Under 2%-8% polypropylene content, an exponential relationship was shown in the ratio of characteristic peak area. This method features fast detection speed and accurate judgment. It can quickly determine whether the polypropylene content in the polyethylene and polypropylene blends is greater than 5% in the tariff inspection. Also, it can determine plastic solid waste attributes and identify whether polyethylene is mixed with a small amount of polypropylene content.

Liver vessel segmentation is fast becoming a key instrument in the diagnosis and surgical planning of liver diseases. In clinical practice, liver vessels are normally manual annotated by clinicians on each slice of CT images, which is extremely laborious. Several deep learning methods exist for liver vessel segmentation, however, promoting the performance of segmentation remains a major challenge due to the large variations and complex structure of liver vessels. Previous methods mainly using existing UNet architecture, but not all features of the encoder are useful for segmentation and some even cause interferences. To overcome this problem, we propose a novel deep neural network for liver vessel segmentation, called LVSNet, which employs special designs to obtain the accurate structure of the liver vessel. Specifically, we design Attention-Guided Concatenation (AGC) module to adaptively select the useful context features from low-level features guided by high-level features. The proposed AGC module focuses on capturing rich complemented information to obtain more details. In addition, we introduce an innovative multi-scale fusion block by constructing hierarchical residual-like connections within one single residual block, which is of great importance for effectively linking the local blood vessel fragments together. Furthermore, we construct a new dataset containing 40 thin thickness cases (0.625 mm) which consist of CT volumes and annotated vessels. To evaluate the effectiveness of the method with minor vessels, we also propose an automatic stratification method to split major and minor liver vessels. Extensive experimental results demonstrate that the proposed LVSNet outperforms previous methods on liver vessel segmentation datasets. Additionally, we conduct a series of ablation studies that comprehensively support the superiority of the underlying concepts.

The real-world high-dimensional data lie on low-dimensional manifolds embedded within the high-dimensional space. Therefore, clustering in high-dimensional spaces is a difficult problem. Subspace-based clustering methods are proposed to project the high dimensional data into a low-dimensional space, and then find clusters in this low-dimensional subspaces of the high dimensional data, instead of finding clusters in the entire feature space. In this work, we propose a subspace clustering method called Sparse Convolution Subspace Clustering (SCSC) which is inspired by Sparse Subspace Clustering (SSC). SSC is to find a sparse representations of a data point in terms of other points while SCSC tries to find a sparse convolutional representations of a data point in terms of other points. A group optimization method based alternating direction method of multipliers (ADMM) is used to solve the sparse convolutional representation problem. It should be pointed out that SSC is a special case of SCSC while the convolution kernel size is set as $1 \times 1$. The experimental results on face data show the effectiveness of the proposed SCSC.

In cloud service systems, customers will report the service issues they have encountered to cloud service providers. Despite many issues can be handled by the support team, sometimes the customer issues can not be easily solved, thus raising customer incidents. Quick troubleshooting of a customer incident is critical. To this end, a customer incident should be assigned to its responsible team accurately in a timely manner. Our industrial experiences show that linking customer incidents with detected system incidents can help the customer incident triage. In particular, our empirical study on 7 real cloud service systems shows that with the additional information about the system incidents (i.e., incident reports generated by system monitors), the triage time of customer incidents can be accelerated 13.1× on average. Based on this observation, in this paper, we propose LinkCM, a learning based approach to automatically link customer incidents to monitor reported system incidents. LinkCM incorporates a novel learning-based model that effectively extracts related information from two resources, and a transfer learning strategy is proposed to help LinkCM achieve better performance without huge amount of data. The experimental results indicate that LinkCM is able to achieve accurate link prediction. Furthermore, case studies are presented to demonstrate how LinkCM can help the customer incident triage procedure in real production cloud service systems.

The sudden outbreak of novel coronavirus 2019 (COVID-19) increased the diagnostic burden of radiologists. In the time of an epidemic crisis, we hope artificial intelligence (AI) to reduce physician workload in regions with the outbreak, and improve the diagnosis accuracy for physicians before they could acquire enough experience with the new disease. In this paper, we present our experience in building and deploying an AI system that automatically analyzes CT images and provides the probability of infection to rapidly detect COVID-19 pneumonia. The proposed system which consists of classification and segmentation will save about 30%–40% of the detection time for physicians and promote the performance of COVID-19 detection. Specifically, working in an interdisciplinary team of over 30 people with medical and/or AI background, geographically distributed in Beijing and Wuhan, we are able to overcome a series of challenges (e.g. data discrepancy, testing time-effectiveness of model, data security, etc.) in this particular situation and deploy the system in four weeks. In addition, since the proposed AI system provides the priority of each CT image with probability of infection, the physicians can confirm and segregate the infected patients in time. Using 1,136 training cases (723 positives for COVID-19) from five hospitals, we are able to achieve a sensitivity of 0.974 and specificity of 0.922 on the test dataset, which included a variety of pulmonary diseases., Highlights • We build a system for detecting COVID-19 infection regions. • We build a dataset of COVID-19. • It is easy to distinguish pneumonia from healthy cases.

The crustal low velocity layers (CLVL) occuring in the middle-lower crust have been found in many regions around the world and have been explained in different ways. In this paper, the compressional wave velocities (VP) of 138 rock samples selected from Qinling and North China were measured under high temperatures (up to 1500 °C) and high pressures (up to 3 GPa) simultaneously. The general feature of the experimental data were discussed and the Vp decrease phenomenon (54 samples involved) were found. Detailed studies, mineral constitutions and textures of experimental products examined by microscope and EPMA, indicated that the dehydration, phase transition and partial melt of water-bearing minerals (hornblende and biotite), are responsible for the Vp decrease. The results suggest that dehydration melting of hydrous minerals might be an important mechanism for CLVL observed in the continental crust of the study area and probably other parts of the world.

The problem of finding a minimum dominating set (MinDS) in a graph is a well known combinatorial optimization problem with wide applications. Due to its NP-hardness, for large and hard instances one must resort to heuristic approaches to obtain good solutions within reasonable time. This paper develops an efficient local search algorithm for MinDS, which has two main ideas. The first one is a novel local search framework, while the second is a construction procedure with inference rules. Our algorithm named FastDS is evaluated on 4 standard benchmarks and 3 massive graphs benchmarks. FastDS obtains the best performance for almost all benchmarks. Particularly, FastDS obtains significantly better solutions than state-of-the-art algorithms on massive graphs.

The mobile news application is a brand-new form of expression of new media. It is the latest form of media carrier in the development of mobile network news reading. With its unique advantages, it provides users with continuous, timely and overall news information, and brings great commercial value to businesses. Based on two management models and the new media characteristics of mobile news app this paper constructs the user stickiness of the news apps influence factors model from different perspectives of product factors and user factors; uses the questionnaire survey method, understands the mobile news apps user experience satisfaction aspects and the mobile news apps present existence questions, takes the existing users of mobile Tencent News app as the research object and conducts empirical research. Through measurement model analysis and structural equation model analysis, it explores the influencing factors of user stickiness.

Characteristics of five subpopulation leukocytes in single-cell levels based on partial principal component analysis coupled with Raman spectroscopy were proposed to recognize the biochemical features of five subpopulation leukocytes. Using wavelet transform, the reconstructed spectra of the low-frequency wavelet coefficients were used to perform multiple principal component analysis based on segmented spectral data wreathing cover at 720–800 cm–1, 840–994 cm–1, and 1010–1070 cm–1 wavenumbers, respectively. Our approach is promising since it enables to establish a better understanding of the underlying molecular difference between the subtypes of leukocytes in a label-free manner and to estimate the source of infection.