In this paper, we investigate a low-power architecture for designs modeled as an Extended Finite State Machine (EFSM). It is based on the general dynamic power management concept, in which the redundant computation can be dynamically disabled to reduce the overall power dissipation. The contribution of this paper is mainly a systematic procedure to identify almost maximal amount of redundant computation in a design given as an EFSM. There are two levels of redundant computation to be exploited--one is based on the machine state information, while the other is based on the transition information. After the extraction of the redundant computation, a low-power architecture using input gating is proposed to synthesize the final circuit. We tested the technique on a design computing a number's modulo inverse. Experimental results show that 31% power reduction can be achieved at the costs of 2% timing penalty and 16% area overhead.

Person Re-Identification has received extensive study in the past few years and achieves impressive progress. Recent outstanding methods extract discriminative features by slicing feature maps of deep neural network into several stripes. Still there have improvement on feature fusion and metric learning strategy which can help promote slice-based methods. In this paper, we propose a novel framework that is end-to-end trainable, called Multi-level Slice-based Network (MSN), to capture features both in different levels and body parts. Our model consists of a dual-branch network architecture, one branch for global feature extraction and the other branch for local ones. Both branches process multi-level features using pyramid feature alike module. By concatenating the global and local features, distinctive features are exploited and properly compared. Also, our proposed method creatively introduces a triplet-center loss to elaborate combined loss function, which helps train the joint-learning network. By demonstrating the comprehensive experiments on the mainstream evaluation datasets including Market-1501, DukeMTMC, CUHK03, it indicates that our proposed model robustly achieves excellent performance and outperforms many of existing approaches. For example, on DukeMTMC dataset in single-query mode, we obtain a great result of Rank-1/mAP =85.9%(+1.0%)/74.2%(+4.7%).

Microblog data contains rich information of real-world events with great commercial values, so microblog-oriented natural language processing (NLP) tasks have grabbed considerable attention of researchers. However, the performance of microblog-oriented Chinese Word Segmentation (CWS) based on deep neural networks (DNNs) is still not satisfying. One critical reason is that the existing microblog-oriented training corpus is inadequate to train effective weight matrices for DNNs. In this paper, we propose a novel active learning method to extend the scale of the training corpus for DNNs. However, due to a large amount of partially overlapped sentences in the microblogs, it is difficult to select samples with high annotation values from raw microblogs during the active learning procedure. To select samples with higher annotation values, parameter λ is introduced to control the number of repeatedly selected samples. Meanwhile, various strategies are adopted to measure the overall annotation values of a sample during the active learning procedure. Experiments on the benchmark datasets of NLPCC 2015 show that our λ-active learning method outperforms the baseline system and the state-of-the-art method. Besides, the results also demonstrate that the performances of the DNNs trained on the extended corpus are significantly improved.

Vehicle detection is challenging in natural traffic scenes because there exist a lot of occlusion. Because of occlusion, detector's training strategy may lead to mismatch between features and labels. As a result, some predicted bounding boxes may shift to surrounding vehicles and lead to lower confidences. These bounding boxes will lead to lower AP value. In this letter, we propose a new approach to address this problem. We calculate the center of visible part of current vehicle based on road information. Then a variable-radius Gaussian weight based method is applied to reweight each anchor box in loss function based on the center of visible part in training time of SSD. The reweighted method has ability to predict higher confidences and more accurate bounding boxes. Besides, the model also has high speed and can be trained end-to-end. Experimental results show that our proposed method outperforms some competitive methods in terms of speed and accuracy.

Broadband ISDN, using asynchronous transfer mode, are expected to carry traffic of different classes, each with its own set of traffic characteristics and performance requirements. To achieve the quality of service in ATM networks, a suitable buffer management scheme is needed. In this paper, we propose a buffer management scheme using a priority service discipline to improve the delay time of delay-sensitive class and the packet loss ratio of loss-sensitive class. The proposed priority scheme requires simple buffer management logic and minor processing overhead. We also analyze the delay time and the packet loss ratio for each class of service. The results indicate that the required buffer size of the proposed priority scheme is reduced and the delay time of each class of service is controlled by a parameter. If the control parameter is appropriately chosen, the quality of service of each class is improved.

This paper proposes a scalable video delivery scheme, named P2PVD, for large-scale video-on-demand applications based on the emerging peer-to-peer structure and characteristic user behaviors. Two types of orders are permitted in P2PVD: reserved and urgent. Reserved orders are encouraged with a lower price policy, which smoothes the network traffic and reduces the server load. The requesting peers use delay-aware dynamic parallel transmission to serve the reserved and urgent orders, and supplying peers employ three priority rules to increase the capacity of P2PVD. Experimental results indicate that P2PVD is scalable.

This study presents several optimization approaches for the MPEG-2/4 Audio Advanced Coding (AAC) Low Complexity (LC) encoding and decoding processes. Considering the power consumption and the peripherals required for consumer electronics, this study adopts the TI OMAP5912 platform for portable devices. An important optimization issue for implementing AAC codec on embedded and mobile devices is to reduce computational complexity and memory consumption. Due to power saving issues, most embedded and mobile systems can only provide very limited computational power and memory resources for the coding process. As a result, modifying and simplifying only one or two blocks is insufficient for optimizing the AAC encoder and enabling it to work well on embedded systems. It is therefore necessary to enhance the computational efficiency of other important modules in the encoding algorithm. This study focuses on optimizing the Temporal Noise Shaping (TNS), Mid/Side (M/S) Stereo, Modified Discrete Cosine Transform (MDCT) and Inverse Quantization (IQ) modules in the encoder and decoder. Furthermore, we also propose an efficient memory reduction approach that provides a satisfactory balance between the reduction of memory usage and the expansion of the encoded files. In the proposed design, both the AAC encoder and decoder are built with fixed-point arithmetic operations and implemented on a DSP processor combined with an ARM-core for peripheral controlling. Experimental results demonstrate that the proposed AAC codec is computationally effective, has low memory consumption, and is suitable for low-cost embedded and mobile applications.

This paper studies a harvested power-oriented simultaneous wireless information and power transfer (SWIPT) scheme over multiple-input multiple-output (MIMO) interference channels in which energy harvesting (EH) circuits exhibit nonlinearity. To maximize the power harvested by all receivers, we propose an algorithm to jointly optimize the transmit beamforming vectors, power splitting (PS) ratios and the receive decoding vectors. As all variables are coupled to some extent, the problem is non-convex and hard to solve. To deal with this non-convex problem, an iterative optimization method is proposed. When two variables are fixed, the third variable is optimized. Specifically, when the transmit beamforming vectors are optimized, the transferred objective function is the sum of several fractional functions. Non-linear sum-of-ratios programming is used to solve the transferred objective function. The convergence and advantage of our proposed scheme compared with traditional EH circuits are validated by simulation results.

Wireless powered communication networks (WPCNs) utilize the wireless energy transfer (WET) technique to facilitate the wireless information transmission (WIT) of nodes. We propose a two-step iterative algorithm to maximize the sum throughput of the users in a MIMO WPCN with discrete signal inputs. Firstly, the optimal solution of a convex power allocation problem can be found given a fixed time allocation; Secondly, a semi closed form solution for the optimal time allocation is obtained when fixing the power allocation matrix. By optimizing the power allocation and time allocation alternately, the two-step algorithm converges to a local optimal point. Simulation results show that the proposed algorithm outperforms the conventional schemes, which consider only Gaussian inputs.