As Deep Neural Networks (DNNs) are evolving in complexity to meet the demands of novel applications, a single device becomes insufficient for training, leading to the emergence of distributed DNN training. However, th...
详细信息
ISBN:
(数字)9798350303582
ISBN:
(纸本)9798350303599
As Deep Neural Networks (DNNs) are evolving in complexity to meet the demands of novel applications, a single device becomes insufficient for training, leading to the emergence of distributed DNN training. However, this evolution exposes a gap in research surrounding security vulnerabilities on model poisoning attacks, especially in model parallel setups, an area that has been scarcely studied. To bridge this gap, we introduce Patronus, an approach that counters model poisoning attacks in distributed DNN training, accommodating both data and model parallelism. With the employment of Loss-aware Credit Evaluation, Patronus scores each participating client. Based on the continuously updated credit, malicious clients are isolated and detected after multiple epochs by Shuffling-based Isolation Mechanism. Additionally, the training system is reinforced by Byzantine Fault-tolerant Aggregation to minimize malicious client impacts. Comprehensive experiments confirm Patronus's superior reliable and efficient performance over the existing methods under attack scenarios.
Failure recovery is one of the most essential problems in Internet of Things (IoT) systems, and the conventional snapshot method is an effective way to solve this problem. However, snapshot methods lack specialized de...
详细信息
ISBN:
(数字)9798331509712
ISBN:
(纸本)9798331509729
Failure recovery is one of the most essential problems in Internet of Things (IoT) systems, and the conventional snapshot method is an effective way to solve this problem. However, snapshot methods lack specialized designs for heterogeneous IoT devices, and when implemented in edge devices, serious system interruptions occur and performance is impacted. To address these problems, a dynamic checkpointing strategy is proposed for IoT systems that consist of heterogeneous devices. Firstly, an anomaly detection network for snapshots (i.e., ADSnet) that combines long short-term memory networks with multilayer convolutional networks is used to learn the multidimensional features of system resource usage. Secondly, ADSnet is tuned during deployment to learn the behaviors of target devices, so that ADSnet can report the anomalies of target devices in the near future. Finally, a dynamic checkpointing strategy is proposed to dynamically create snapshots on the basis of the anomaly detection results. The experimental results show that the proposed ADSnet achieves 97.73% accuracy in detecting anomalies in the target device; furthermore, our proposed dynamic checkpointing strategy reduces 25.4% snapshots than that created by the recently proposed ResCheck.
NASA has committed to open-source science that enables Earth observation data transparency, inclusivity, accessibility, and reproducibility - all fundamental to the pace and quality of scientific progress. We have emb...
详细信息
ISBN:
(纸本)9798350320107
NASA has committed to open-source science that enables Earth observation data transparency, inclusivity, accessibility, and reproducibility - all fundamental to the pace and quality of scientific progress. We have embraced this vision by producing standard InSAR science products that are freely available to the public through NASA Data Active Archive Centers (DAACs) and are generated using state-of-the-art open-source and openly-developed methods. The Advanced Rapid image Analysis (ARIA) project's Sentinel-1 Geocoded Unwrapped Phase product (ARIA-S1-GUNW) is a 90 meter InSAR product that spans major, land-based fault systems, the US Coasts, and active volcanic regions through the complete Sentinel-1 record. The products enable the measurement of centimeter-scale surface displacement with applications across the solid earth, hydrology, and sea-level disciplines. The ARIA-S1-GUNW also enables rapid response mapping of surface motion after earthquakes, landslides, and subsidence. The ARIA-S1-GUNW products are freely available through the Alaska Satellite Facility (ASF) DAAC. In the last year, we have successfully grown the archive to over 1.1 million products, a 6 fold increase, through NASA ACCESS by improving our processing workflow and leveraging HyP3, an AWS-based cloud processing environment. We are continuing to partner with researchers to generate more products over relevant areas of scientific interest. All the processing software and cloud infrastructure are open-source to ensure reproducibility and enable other scientists to modify, improve upon, and scale their own cloud workflows for related InSAR analyses. We have, in parallel, developed and supported open-source, well-documented tools to further streamline time-series analysis from the ARIA-S1-GUNW into deformation analysis workflows.
This article presents a GPU-accelerated software design of the recently proposed model of Slanted Stixels, which represents the geometric and semantic information of a scene in a compact and accurate way. We reformula...
详细信息
This article presents a GPU-accelerated software design of the recently proposed model of Slanted Stixels, which represents the geometric and semantic information of a scene in a compact and accurate way. We reformulate the measurement depth model to reduce the computational complexity of the algorithm, relying on the confidence of the depth estimation and the identification of invalid values to handle outliers. The proposed massively parallel scheme and data layout for the irregular computation pattern that corresponds to a Dynamic Programming paradigm is described and carefully analyzed in performance terms. Performance is shown to scale gracefully on current generation embedded GPUs. We assess the proposed methods in terms of semantic and geometric accuracy as well as run-time performance on three publicly available benchmark datasets. Our approach achieves real-time performance with high accuracy for 2048 x 1024 image sizes and 4 x 4 Stixel resolution on the low-power embedded GPU of an NVIDIA Tegra Xavier.
The particle Markov-chain Monte Carlo (PMCMC) method is a stochastic algorithm that combines Particle Filters (PFs) and Markov-chain Monte Carlo (MCMC) techniques. This approach is widely used in Bayesian inference fo...
详细信息
ISBN:
(数字)9798350350920
ISBN:
(纸本)9798350350937
The particle Markov-chain Monte Carlo (PMCMC) method is a stochastic algorithm that combines Particle Filters (PFs) and Markov-chain Monte Carlo (MCMC) techniques. This approach is widely used in Bayesian inference for high-dimensional state spaces and nonlinear, non-Gaussian dynamic systems. However, current PMCMC accelerators face significant challenges due to their intensive computational complexity and the intricate particle routing, limiting their application in real-time scenarios. To address these challenges, we propose a novel distributed PMCMC method that leverages parallel computing to enhance hardware execution speed. Additionally, our method introduces a particle exchange scheme that not only resolves the accuracy issues caused by particle routing in distributed PMCMC but also achieves faster computing speed. Our design is implemented on a Xilinx Kintex-7 xc7k480t FPGA device. Experimental results demonstrate that our accelerator is nearly 65 ×faster than CPU performance, and provides speedups up to 5× compared to existing FPGA-based accelerators.
Accurately mapping the surface rivers is important in ecological environment monitoring and disaster prevention. The development of remote sensing technology and computer vision greatly improves the efficiency of this...
详细信息
Neural Radiance Field (NeRF) has received widespread attention for its photo-realistic novel view synthesis quality. Current methods mainly represent the scene based on point sampling of ray casting, ignoring the infl...
Neural Radiance Field (NeRF) has received widespread attention for its photo-realistic novel view synthesis quality. Current methods mainly represent the scene based on point sampling of ray casting, ignoring the influence of the observed area changing with distance. In addition, The current sampling strategies are all focused on the distribution of sampling points on the ray, without paying attention to the sampling of the ray. We found that the current ray sampling strategy for scenes with the camera moving forward severely reduces the convergence speed. In this work, we extend the point representation to area representation by using relative positional encoding, and propose a ray sampling strategy that is suitable for camera trajectory moving forward. We validated the effectiveness of our method on multiple public datasets.
This study conducts an in-depth evaluation of imaging algorithms and software and hardware architectures to meet the capability requirements of real-time image acquisition systems, such as spaceborne and airborne synt...
详细信息
This study conducts an in-depth evaluation of imaging algorithms and software and hardware architectures to meet the capability requirements of real-time image acquisition systems, such as spaceborne and airborne synthetic aperture radar (SAR) systems. By analysing the principles and models of SAR imaging, this research creatively puts forward the fully parallelprocessing architecture for the back projection (BP) algorithm based on Field-Programmable Gate Array (FPGA). The processing time consumption has significant advantages compared with existing methods. This article describes the BP imaging algorithm, which stands out with its high processing accuracy and two-dimensional decoupling of distance and azimuth, and analyses the algorithmic flow, operation, and storage requirements. The algorithm is divided into five core operations: range pulse compression, upsampling, oblique distance calculation, data reading, and phase accumulation. The architecture and optimisation of the algorithm are presented, and the optimisation methods are described in detail from the perspective of algorithm flow, fixed-point operation, parallelprocessing, and distributed storage. Next, the maximum resource utilisation rate of the hardware platform in this study is found to be more than 80%, the system power consumption is 21.073 W, and the processing time efficiency is better than designs with other FPGA, DSP, GPU, and CPU. Finally, the correctness of the processing results is verified using actual data. The experimental results showed that 1.1 s were required to generate an image with a size of 900 x 900 pixels at a 200 MHz clock rate. This technology can solve the multi-mode, multi-resolution, and multi-geometry signal processing problems in an integrated manner, thus laying a foundation for the development of a new, high-performance, SAR system for real-time imaging processing.
This dissertation presents our research and development works to address the challenges for delivering effective, scalable, and high performance methods and systems for managing, and querying complex spatial and spati...
This dissertation presents our research and development works to address the challenges for delivering effective, scalable, and high performance methods and systems for managing, and querying complex spatial and spatio-temporal big data at multiple dimensions. This is driven mainly by big data problems emerging from geospatial applications, medical image analysis, satellite imaging, weather analytics, and collaborative spatial data collection. Managing and analyzing such data poses major challenges, including explosion of data volume, an unprecedented rate of data generation, and high complexity of handling spatial and/or temporal dynamics. Traditional spatial data processing algorithms and data structures have been developed mostly for linear processing. Despite the need, these are not optimized for modern requirements of big data processing. To this end, we present strategies for adapting spatial and spatio-temporal principles and methods to the parallel and distributed domain. In particular, we approached the problem from three different aspects. In distributed domain, recent attempts have proposed Hadoop based strategies to accelerate big spatial data processing. Hadoop based systems, however, rely on disk IO, fairly limiting system performance. Apache Spark, on the other hand, is a widely popular framework, based on similar distributed principles, providing iterative in-memory processing. Despite their efficiency, Spark based systems are oblivious to spatial characteristics and provide limited memory optimizations catering to spatial data. To address these, we designed and implemented SparkGIS; which not only provid in-memory spatial querying capabilities but also introduce a novel resource-aware dynamic query re-writing mechanism to efficiently manage distributed resources for large spatial workloads. Spatial processing is inherently compute intensive due to its complexity and multi-dimensional properties. While SparkGIS addresses in-memory aspects of spatial
In sentence similarity research methods, sentence similarity is often calculated from semantic aspects, however, the influence of other features is ignored. For example, the influence of sentence syntactic structure a...
In sentence similarity research methods, sentence similarity is often calculated from semantic aspects, however, the influence of other features is ignored. For example, the influence of sentence syntactic structure and word order on sentence similarity is not considered. Therefore, This paper proposes a method to measure the degree of similarity between sentences based on a multi-feature fusion of semantics, syntactic similarity and word order similarity, respectively. For the aspect of semantic similarity, in this paper, we first preprocess the sentence and obtain the word vectors, and then obtains sentence vector features through neural networks, so as to calculate the sentence semantic similarity; Finally, for the aspect of word order similarity, word order is considered as the auxiliary feature information to calculate sentence similarity since word order has a smaller effect on sentence similarity compared to the above information. The experimental results show that the proposed model has an accuracy of 77.4 % and an F1 value of 84.3% on the MRPC dataset.
暂无评论