Journal of Computer Science and Technology最新文献

Event-based computation has recently gained increasing research interest for applications of vision recognition due to its intrinsic advantages on efficiency and speed. However, the existing event-based models for vision recognition are faced with several issues, such as large network complexity and expensive training cost. In this paper, we propose an improved multi-liquid state machine (M-LSM) method for high-performance vision recognition. Specifically, we introduce two methods, namely multi-state fusion and multi-liquid search, to optimize the liquid state machine (LSM). Multistate fusion by sampling the liquid state at multiple timesteps could reserve richer spatiotemporal information. We adapt network architecture search (NAS) to find the potential optimal architecture of the multi-liquid state machine. We also train the M-LSM through an unsupervised learning rule spike-timing dependent plasticity (STDP). Our M-LSM is evaluated on two event-based datasets and demonstrates state-of-the-art recognition performance with superior advantages on network complexity and training cost.

The query model (or black-box model) has attracted much attention from the communities of both classical and quantum computing. Usually, quantum advantages are revealed by presenting a quantum algorithm that has a better query complexity than its classical counterpart. In the history of quantum algorithms, the Deutsch algorithm and the Deutsch-Jozsa algorithm play a fundamental role and both are exact one-query quantum algorithms. This leads us to consider the problem: what functions can be computed by exact one-query quantum algorithms? This problem has been addressed in the literature for total Boolean functions and symmetric partial Boolean functions, but is still open for general partial Boolean functions. Thus, in this paper, we continue to characterize the computational power of exact one-query quantum algorithms for general partial Boolean functions. First, we present several necessary and sufficient conditions for a partial Boolean function to be computed by exact one-query quantum algorithms. Second, inspired by these conditions, we discover some new representative functions that can be computed by exact one-query quantum algorithms but have an essential difference from the already known ones. Specially, it is worth pointing out that before our work, the known functions that can be computed by exact one-query quantum algorithms are all symmetric functions and the quantum algorithm used is essentially the Deutsch-Jozsa algorithm, whereas the functions discovered in this paper are generally asymmetric and new algorithms to compute these functions are required. Thus, this expands the class of functions that can be computed by exact one-query quantum algorithms.

The goal of privacy-preserving social graph release is to protect individual privacy while preserving data utility. Community structure, which is an important global pattern of nodes, is a crucial data utility as it is fundamental to many graph analysis tasks. Yet, most existing methods with differential privacy (DP) commonly fall into edge-DP to sacrifice security in exchange for utility. Moreover, they reconstruct graphs from the local feature-extraction of nodes, resulting in poor community preservation. Motivated by this, we develop PrivCom, a strict node-DP graph release algorithm to maximize the utility on the community structure while maintaining a higher level of privacy. In this algorithm, to reduce the huge sensitivity, we devise a Katz index based private graph feature extraction method, which can capture global graph structure features while greatly reducing the global sensitivity via a sensitivity regulation strategy. Yet, under the condition that the sensitivity is fixed, the feature captured by the Katz index, which is presented in matrix form, requires privacy budget splits. As a result, plenty of noise is injected, mitigating global structural utility. To bridge this gap, we design a private eigenvector estimation method, which yields noisy eigenvectors from extracted low-dimensional vectors. Then, a dynamic privacy budget allocation method with provable utility guarantees is developed to preserve the inherent relationship between eigenvalues and eigenvectors, so that the utility of the generated noise Katz matrix is well maintained. Finally, we reconstruct the synthetic graph via calculating its Laplacian with the noisy Katz matrix. Experimental results confirm our theoretical findings and the efficacy of PrivCom.

Skeletons of characters provide vital information to support a variety of tasks, e.g., optical character recognition, image restoration, stroke segmentation and extraction, and style learning and transfer. However, automatically skeletonizing Chinese characters poses a steep computational challenge due to the large volume of Chinese characters and their versatile styles, for which traditional image analysis approaches are error-prone and fragile. Current deep learning based approach requires a heavy amount of manual labeling efforts, which imposes serious limitations on the precision, robustness, scalability and generalizability of an algorithm to solve a specific problem. To tackle the above challenge, this paper introduces a novel three-staged deep generative model developed as an image-to-image translation approach, which significantly reduces the model’s demand for labeled training samples. The new model is built upon an improved G-net, an enhanced X-net, and a newly proposed F-net. As compellingly demonstrated by comprehensive experimental results, the new model is able to iteratively extract skeletons of Chinese characters in versatile styles with a high quality, which noticeably outperforms two state-of-the-art peer deep learning methods and a classical thinning algorithm in terms of F-measure, Hausdorff distance, and average Hausdorff distance.

The emerging mobile robot industry has spurred a flurry of interest in solving the simultaneous localization and mapping (SLAM) problem. However, existing SLAM platforms have difficulty in meeting the real-time and low-power requirements imposed by mobile systems. Though specialized hardware is promising with regard to achieving high performance and lowering the power, designing an efficient accelerator for SLAM is severely hindered by a wide variety of SLAM algorithms. Based on our detailed analysis of representative SLAM algorithms, we observe that SLAM algorithms advance two challenges for designing efficient hardware accelerators: the large number of computational primitives and irregular control flows. To address these two challenges, we propose a hardware accelerator that features composable computation units classified as the matrix, vector, scalar, and control units. In addition, we design a hierarchical instruction set for coping with a broad range of SLAM algorithms with irregular control flows. Experimental results show that, compared against an Intel x86 processor, on average, our accelerator with the area of 7.41 mm² achieves 10.52x and 112.62x better performance and energy savings, respectively, across different datasets. Compared against a more energy-efficient ARM Cortex processor, our accelerator still achieves 33.03x and 62.64x better performance and energy savings, respectively.