Conventional error correction codes (ECCs), such as the commonly used BCH code, have become increasingly inadequate for solid state drives (SSDs) capacity of NAND flash memory continues to increase and its reliability degrade. It is highly desirable deploy a much more powerful ECC, low-density parity-check (LDPC) significantly improve SSDs. Although LDPC code has had success in commercial hard disk drives, fully exploit capability SSDs demands unconventional fine-grained sensing, leading an...

With the potential to increase road safety and provide economic benefits, intelligent vehicles have elicited a significant amount of interest from both academics industry. A robust reliable vehicle detection tracking system is one key modules for perceive surrounding environment. The millimeter-wave radar monocular camera are two vehicular sensors commonly used tracking. Despite their advantages, drawbacks these make them insufficient when separately. Thus, fusion considered as an efficient...

Mainstream object detectors based on the fully convolutional network has achieved impressive performance. While most of them still need a hand-designed non-maximum suppression (NMS) post-processing, which impedes end-to-end training. In this paper, we give analysis discarding NMS, where results reveal that proper label assignment plays crucial role. To end, for detectors, introduce Prediction-aware One-To-One (POTO) classification to enable detection, obtains comparable performance with NMS....

LiDAR and camera are two critical sensors that can provide complementary information for accurate 3D object detection. Most works devoted to improving the detection performance of fusion models on clean well-collected datasets. However, collected point clouds images in real scenarios may be corrupted various degrees due potential sensor malfunctions, which greatly affects robustness model poses a threat safe deployment. In this paper, we first analyze shortcomings most detectors, rely mainly...

Because of its high storage density with superior scalability, low integration cost and reasonably access speed, spin-torque transfer random memory (STT RAM) appears to have a promising potential replace SRAM as last-level on-chip cache (e.g., L2 or L3 cache) for microprocessors. Due unique operational characteristics device magnetic tunneling junction (MTJ), STT RAM is inherently subject write latency versus read tradeoff that determined by the cell size. This paper first quantitatively...

Current state-of-the-art approaches for spatio-temporal action detection have achieved impressive results but remain unsatisfactory temporal extent detection. The main reason comes from that, there are some ambiguous states similar to the real actions which may be treated as target even by a well trained network. In this paper, we define these samples "transitional states", and propose Transition-Aware Context Network (TACNet) distinguish transitional states. proposed TACNet includes two...

Source code summaries are important for program comprehension and maintenance. However, there plenty of programs with missing, outdated, or mismatched summaries. Recently, deep learning techniques have been exploited to automatically generate given snippets. To achieve a profound understanding how far we from solving this problem provide suggestions future research, in paper, conduct systematic in-depth analysis 5 state-of-the-art neural summarization models on 6 widely used BLEU variants, 4...

Over the past decades, there has been significant research effort dedicated to development of autonomous vehicles and advanced driver assistance systems. The driving control system, which is responsible for trajectory tracking safety, one most important technologies vehicles. This paper describes design including both longitudinal lateral controllers, Kuafu-II vehicle. Compared with previous researches that inevitably require a large amount parameters, presented system in this integrates...

3D human action representation and recognition are important issues in many multimedia applications. While latent state approaches have been widely used for modeling, previous works assume the states of actions single attribute. This assumption is inaccurate representing structures complex actions. In this paper, we propose that composite attributes introduce a novel structure (CLS) model to represent recognize with skeleton sequences. A modeled hierarchical graph, which represents sequence...

With the deployment of smart sensors and advancements in communication technologies, big data analytics have become vastly popular grid domain, informing stakeholders best power utilization strategy. However, these power-related are stored owned by different parties. For example, consumption numerous transformer stations across cities; mobility population, which important indicators consumption, held mobile companies. Direct sharing might compromise party benefits, individual privacy even...

Today, an increasing number of applications and services are being hosted by large-scale data centers. The massive irregular load surges challenge center power infrastructures. As a result, mismatching between supply demand has emerged as crucial issue in modern centers which either under-provisioned or powered intermittent sources. Recent proposals have employed energy storage devices such the uninterruptible (UPS) systems to address this issue. However, current approaches lack capacity...

A multilevel per cell (MLC) technique significantly improves the storage density, but also poses serious data integrity challenge for NAND flash memory. This consequently makes low-density parity-check (LDPC) code and soft-decision memory sensing become indispensable in next-generation flash-based solid-state devices. However, use of LDPC codes inevitably increases read latency and, hence, degrades speed performance. Motivated by observation intracell unbalanced bit error probability...

This paper addresses the issue of frontal object perception in real-world traffic scenarios. Accurate and real-time plays a key role Advanced Driver Assistance Systems (ADAS) Intelligent Vehicles (IV). However, perceiving complex environments, which contain multiple classes on-road objects with various visual appearances from different viewpoints partial observations, is still challenging task. In this paper, system fusing millimeter-wave (MMW) radar monocular camera proposed. Firstly,...

In this paper, a novel method, relying on Markov Transition Fields (MTF) and deep learning, is proposed to classify the vibration-events measure vibration-frequency, for <inline-formula xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink"> <tex-math notation="LaTeX">$ {\varphi } $ </tex-math></inline-formula> -OTDR based fiber-optic distributed vibration sensor. The normalized time series of signal detected by notation="LaTeX">${\varphi }$ are converted...

In heterogeneous wireless networks, massive mobile terminals randomly generate a large number of computation tasks (payloads). How to better manage these located in networks achieve acceptable quality service (QoS) such as latency minimization, energy consumption minimization is crucial. A multi-access edge computing (MEC) server can be leveraged execute the offloaded payloads generated from owing its powerful processing power and location proximity features. However, an MEC cannot tackle...

Depthwise separable convolution (DSC) has become one of the essential structures for lightweight convolutional neural networks. Nevertheless, its hardware architecture not received much attention. Several previous designs incur either high off-chip memory traffic or large on-chip usage, and hence have deficiency in terms efficiency as well performance. This paper proposes two efficient dynamic design techniques, i.e. adaptive row-based dataflow scheduling computation mapping, to achieve a...

Code summarization aims to generate concise natural language descriptions of source code, which can help improve program comprehension and maintenance. Recent studies show that syntactic structural information extracted from abstract syntax trees (ASTs) is conducive summary generation. However, existing approaches fail fully capture the rich in ASTs because large size/depth ASTs. In this paper, we propose a novel model CAST hierarchically splits reconstructs First, split AST into set...

Abstract Internet of Things devices generate a large number heterogeneous workloads in real-time that require specific application to tackle, and the inability communicate between communication base stations due complex scenarios is thorny issue. Service caching play key role managing specific-request workload from devices, unmanned aerial vehicles with computation functions can effectively solve barrier ground stations. In addition, joint optimization offloading service cache placement...

Editor's note:From a system architecture perspective, 3D technology can satisfy the high memory bandwidth demands that future multicore/manycore architectures require. This article presents DRAM design and potential for using stacking both L2 cache main in multicore architecture.—Yuan Xie, Pennsylvania State University

Inspired by the cooperation between cone cell and rod in human vision system, this paper presents a novel method for fusing millimeter wave (MMW) radar monocular sensor information on-road obstacle detection. The MMW performs similarly as to indicate potential target region view, while camera plays same role that of obtain precise contour allow tracking over frames. Using real-world data collected from vehicles, experimental evaluations have shown that, our proposed system could detect...

Efficient real-time disparity estimation is critical for the application of stereo vision systems in various areas. Recently, network based on coarse-to-fine method has largely relieved memory constraints and speed limitations large-scale models. Nevertheless, all previous designs employ constant offsets three or more stages to progressively refine coarse map, still resulting unsatisfactory computation accuracy inference time when deployed mobile devices. This paper claims that matching...

As one promising non-volatile memory technology, magnetoresistive RAM (MRAM) based on magnetic tunneling junctions (MTJs) has recently attracted much attention. However, latest device research discovered that, in order to maintain sufficient MTJ write margin prevent breakdown, MTJs will be subject unconventionally high random error rates (e.g., 10-3 and above) as cell size is being scaled down. This new discovery seriously threatens the scalability of MRAM, material/device community actively...