Search

IoT time series data is an important form of big data. How to improve the efficiency of storage system is crucial for IoT time series database to store and manage massive IoT time series data from various IoT devices. Mixing NVM and SSD is an effective method to improve the I/O performance of storage systems. However, there are great differences between HDD and NVM or SSD. As a result, NVM and SSD cannot be directly used in the current time series database effectively. We design an IoT time series database with an embedded engine in storage device drivers for the hybrid solid-state storage system consisting of NVM and SSD. The I/O software stack of storing and managing IoT time series data can be shortened to improve the efficiency. Based upon the intrinsic characteristics of IoT time series data and different features of NVM and SSD, a redundancy elimination and compression fusion strategy, a hierarchical management strategy, and a heterogeneous time series data index are designed to improve the efficiency. Finally, a prototype of embedded IoT time series database named TS-NSM is implemented, and YCSB-TS is used to measure the IOPS. The results show that TS-NSM can improve the write IOPS up to 243.6 times and 174.3 times, respectively, compared with InfluxDB and OpenTSDB, and improve the read IOPS up to 10.1 times and 14.4 times, respectively.

How to improve the I/O speed of storage systems is crucial for the ability of vehicle big data to meet the demand of vehicular social networks. Non-volatile memory (NVM) is an effective method for improving the performance of storage systems. Meanwhile, metadata management efficiency has a significant impact on the I/O performance of storage systems. But the current metadata management strategy is not efficient and will decrease the endurance of NVM devices. We design a new file system with a distributed metadata management strategy for hybrid NVM consisting of NVDIMMs and PCIe NVM devices. Based upon the intrinsic characteristics of two types of NVM devices and different attributes in the metadata, the attributes are divided, reorganized, and distributed between the two types of NVM devices based on the access frequency. The basic attributes with high access frequency are stored in NVDIMMs and managed by the CPU using a hierarchical algorithm to ensure the efficiency of management. The extended attributes with low access frequency are converted to multiple key-value pairs to improve the scalability of management and are stored in the PCIe NVM devices. Then, the embedded processor in the PCIe NVM devices is used to manage these extended attributes in the form of key-value pairs. This can reduce the transfer between the CPU and the PCIe NVM devices and decrease the write amplification of the PCIe NVM devices as well as improve the efficiency of metadata management. Thus, the metadata management tasks can be distributed between the CPU and the embedded processor in the PCIe NVM devices to reduce the time overhead of metadata management. Finally, a prototype named XPMFS is implemented based on PMFS and PMBD. Fio and Filebench are used to test its performance against PMFS and Ext4-DAX on PMBD. For Ext4-DAX on PMBD, the results demonstrate that the XMPFS can reduce the write and read time overhead by 52% and 49%, respectively, improving the I/O performance by 1.5 times under real workloads. The I/O performance of XPMFS is only slightly inferior to that of PMFS, but it avoids the storage capacity limitation of PMFS and aims for hybrid NVM.

Long short-term memory (LSTM) is an important model for sequential data processing. However, large amounts of matrix computations in LSTM unit seriously aggravate the training when model grows larger and deeper as well as more data become available. In this work, we propose an efficient distributed duration-aware LSTM(D-LSTM) for large scale sequential data analysis. We improve LSTM’s training performance from two aspects. First, the duration of sequence item is explored in order to design a computationally efficient cell, called duration-aware LSTM(D-LSTM) unit. With an additional mask gate, the D-LSTM cell is able to perceive the duration of sequence item and adopt an adaptive memory update accordingly. Secondly, on the basis of D-LSTM unit, a novel distributed training algorithm is proposed, where D-LSTM network is divided logically and multiple distributed neurons are introduced to perform the easier and concurrent linear calculations in parallel. Different from the physical division in model parallelism, the logical split based on hidden neurons can greatly reduce the communication overhead which is a major bottleneck in distributed training. We evaluate the effectiveness of the proposed method on two video datasets. The experimental results shown our distributed D-LSTM greatly reduces the training time and can improve the training efficiency for large scale sequence analysis.

The non-volatile memory (NVM) device is a useful way to solve the memory wall in computers. However, the current I/O software stack in operating systems becomes a performance bottleneck for applications based on NVM devices, especially for key&ndash, value stores. We analyzed the characteristics of key&ndash, value stores and NVM devices and designed a new embedded key&ndash, value store for an NVM device simulator named PMEKV. The embedded processor in NVM devices was used to manage key&ndash, value pairs to reduce the data transfer between NVM devices and key&ndash, value applications. Meanwhile, it also cut down the data copy between the user space and the kernel space in the operating system to alleviate the I/O software stacks on the efficiency of key&ndash, value stores. The architecture, data layout, management strategy, new interface and log strategy of PMEKV are given. Finally, a prototype of PMEKV was implemented based on PMEM. We used YCSB to test and compare it with Redis, MongDB, and Memcache. Meanwhile, the Redis for PMEM named PMEM-Redis and PMEM-KV were also used to test and compared with PMEKV. The results show that PMEKV had the advantage of throughput and adaptability compared with the current key&ndash, value stores.

Long short-term memory (LSTM) network is an effective model architecture for deep learning approaches to sequence modeling tasks. However, the current LSTMs can't use the property of sequential data when dealing with the sequence components, which last for a certain period of time. This may make the model unable to benefit from the inherent characteristics of time series and result in poor performance as well as lower efficiency. In this paper, we present a novel adaptive LSTM for durative sequential data which exploits the temporal continuance of the input data in designing a new LSTM unit. By adding a new mask gate and maintaining span, the cell's memory update is not only determined by the input data but also affected by its duration. An adaptive memory update method is proposed according to the change of the sequence input at each time step. This breaks the limitation that the cells calculate the cell state and hidden output for each input always in a unified manner, making the model more suitable for processing the sequences with continuous data. The experimental results on various sequence training tasks show that under the same iteration epochs, the proposed method can achieve higher accuracy, but need relatively less training time compared with the standard LSTM architecture.

The number of hidden neurons in recurrent neural networks (RNNs) greatly affects the model accuracy. When fewer neurons are used, less training time is involved and the accuracy is in a lower degree. Once we increase the number of hidden units, the performance will be encouraged but the training overhead will enhance exponentially. In this work, we propose a novel active propagation training strategy for distributed RNN, which is inspired by the information transmission in biological neuron system. With the aim to improve the training efficiency of RNN, the proposed active propagation avoids the traditional fixed number of neurons in training. Instead, when training the distributed RNN, the autonomous neuron dynamically activates new neurons and propagates the accumulated training information to them. The neurons not only learn from the training data but also from other neurons. Different from a fixed large number of neurons, the valid neurons are increased gradually and the model performance is promoted accordingly. A prototype of active propagation RNN (APRNN) is implemented on Spark and multiple evaluations are carried out on language modeling tasks. The results show that compared with the fixed-neurons training, APRNN can reduce the training time while keeping a comparable accuracy. The improvement is more significant on large scale training corpora.

The Non-Volatile Memory (NVM) device has higher read and write speed compared with disks and SSDs, and it is the hotspot to improve the efficiency of computer system. But it also makes the I/O software stack become the bottleneck of application's efficiency instead of storage devices. We analyse I/O software stack when the application accesses the NVM device and design a direct hybrid NVM file system which can bypass the VFS to reduce the extra time overhead of accessing the NVM device for the application. The application oriented direct file system interface is design to allow the application directly access the NVM device which reducing the response time and improving the I/O performance of applications. On this basis, we implement the prototype of DNVMCFS and use Fio to test its performance comparing with PMFS, NVMCFS and Ext4 on PMBD respectively. The results show that DNVMCFS can improve the I/O performance by 42.3% compared with PMFS, 50% compared with NVMCFS and 53.7% compared with Ext4 on PMBD. Meanwhile, DNVMCFS can also increase the IOPS where Filebench is used to simulate the real application test. The results show that DNMVCFS raises the I/O performance and IOPS by 25%. These results verify that DNVMCFS provides higher I/O performance and better adaptability.

RNNLM (Recurrent Neural Network Language Model) can save the historical information of the training dataset by the last hidden layer and can also as input for training. It has become an interesting topic in the field of Natural Language Processing research. However, the immense training time overhead is a big problem. The large output layer, hidden layer, last hidden layer and the connections among them will generate enormous matrix in training. It is the main facts to influence the efficiency and scalability. At the same time, output layer class and small hidden layer should decrease the accuracy of RNNLM. In general, the lack of parallel for artificial neuron is main reason for these. We change the structure of RNNLM and design the new large-scale RNNLM by the center of distributed artificial neurons in hidden layer to stimulate the parallel characteristic of biological neuron system. Meanwhile, we change training method, and present the coordination strategy for distributed neuron. At last, the prototype of new large-scale RNNLM system is implemented based on Spark. The testing and analysis results show that the training time overhead is far less than the growth rate of the distributed neuron in hidden layer and size of training dataset. These results show our large-scale RNNLM system has efficiency and scalability advantage.

Semantic categorization for the complex videos is an ambiguous task. The semi-supervised learning method based on hypergraph model can achieve multi-semantics labels, but a hypergraph model is sensitive to the radius parameter when it is constructed and the number of vertices belonging to a hyperedge is fixed. In this paper, a semi-supervised incremental learning method based on adaptive probabilistic hypergraph for video semantic detection is presented. In the probabilistic hypergraph modeling, a formula is presented as a measurement to adaptively decide whether a vertex is belonged to a hyperedge or not. The model has high robustness and can overcome the defect of fixed number of vertices belonging to the same hyperedge in the traditional probabilistic hypergraph model. In the semi-supervised incremental learning process, a threshold is defined, which is used to judge whether unlabeled sample can be added into the modeling, in order that the model learning result for unlabeled samples has high certainty. The experimental results show that our method can improve the model generalization ability and utilize the unlabeled samples effectively. In the aspects of recall rate and precision rate for semantic video concept detection from complex videos, our proposed method outperforms the compared methods.

The storage device based on Nonvolatile Memory (NVM devices) has high read/write speed and embedded processor. It is a useful way to improve the efficiency of Key-Value (KV) application. However it still has some limitations such as limited capacity, poorer computing power compared with CPU, and complex I/O system software. Thus it is not an effective way to construct KV storage system with NVM devices directly. We analyze the characteristics of NVM devices and demands of KV application to design the matrix KV storage system based on NVM Devices. The group collaboration management based on Bloomfilter, intragroup optimization based on competition, embedded KV management based on B+-tree, and the new interface of KV storage system are presented. Then, the embedded processor in the NVM device and CPU can be comprehensively utilized to construct a matrix KV pair management system. It can improve the storage and management efficiency of massive KV pairs, and it can also support the efficient execution of KV applications. A prototype is implemented named MKVS (the matrix KV storage system based on NVM devices) to test with YCSB (Yahoo! Cloud System Benchmark) and to compare with the current in-memory KV store. The results show that MKVS can improve the throughput by 5.98 times, and reduce the 99.7% read latency and 77.2% write latency.

The K-nearest neighbour classifier is very effective and simple non-parametric technique in pattern classification, however, it only considers the distance closeness, but not the geometricalplacement of the k neighbors. Also, its classification performance is highly influenced by the neighborhood size k and existing outliers. In this paper, we propose a new local mean based k-harmonic nearest centroid neighbor (LMKHNCN) classifier in orderto consider both distance-based proximity, as well as spatial distribution of k neighbors. In our method, firstly the k nearest centroid neighbors in each class are found which are used to find k different local mean vectors, and then employed to compute their harmonic mean distance to the query sample. Lastly, the query sample is assigned to the class with minimum harmonic mean distance. The experimental results based on twenty-six real-world datasets shows that the proposed LMKHNCN classifier achieves lower error rates, particularly in small sample-size situations, and that it is less sensitive to parameter k when compared to therelated four KNN-based classifiers.

Facial expression recognition technology plays an important role in research areas such as psychological studies, image understanding and virtual reality etc. In order to achieve subject-independent facial expression recognition and obtain robustness against illumination variety and image deformation, facial expression recognition methods based on Gabor wavelet transformation and elastic templates matching are presented in this paper. First given a still image containing facial expression information, preprocessors are executed which include gray and scale normalization. Secondly, Gabor wavelet filters are adopted to extract expression features. Then the elastic graph for expression features is constructed. Finally, elastic templates matching algorithm and K-nearest neighbors classifier are used to recognize facial expression. Experiments show that expression features can be extracted effectively by Gabor wavelet transformation, which is insensitive to illumination variety and individual difference, and high recognition rate can be obtained using elastic templates matching algorithm, which is subject-independent.

SAN-level buffer cache is an important factor in improving the efficiency of the storage area network (SAN). In this paper, we analyzed the SAN-level access pattern characterization, and designed a new hot spot selection algorithm called maximal access times and oldest access first select (MOFS) and minimal access times and oldest access first eliminate (MOFE) for SAN-level buffer cache. The line size for the hot spot is larger than the line size implemented in disk array caches. The algorithm calls in the block with the highest number of access times and oldest access to SAN-level buffer cache, and eliminates the block with the minimal access times and oldest access from the accessed block list. The algorithm uses the self-adapt mechanism to change the parameter’s algorithm value dynamically. We implemented a virtual address switch in the SAN virtualization system to collect the access request information. Base on this we implemented the hot spot selection algorithm to select a block and send it to the SAN-level buffer cache. Lastly we evaluated the MOFS and MOFE algorithm and proved that this algorithm realizes a low call in ratio and high hit ratios in the SAN-level buffer cache and the self-adapt mechanism makes the MOFS and MOFE work efficiently with different types of workloads.

In this paper, we present a special model G-GRID for graphic service grid system. The main function of each layer in the model is described in detail, the service description and system architecture are defined as well. Then we propose the key algorithms of services discovery, management and schedule of the grid system. Some implementation interfaces of the graphic service grid system are provided in the end. Then prove the validity of graphic service grid system.

The speaker often wants to express his own views boldly in some oral argument cases in the virtual community. However, he hopes that others could not recognize him. For this purpose, we adopted the voice modification technique to meet these requirements. In this paper, we presented a method of rotating the Pseudo Log Area Ration (PLAR) curve to modify the representation of the speech based on the experiments of the relation between PLAR coefficients and personal information. We also deduced a series of equations to testify the accuracy of this method. Using this method we developed a real-time voice modification system based on LP model, and PSOLA method was also added to this system to produce a better effectiveness. Experiments show that our system has good property of real-time and satisfactory auditory effectiveness and it has successfully been implemented in the virtual community.