This paper will use the characteristics of decentralized trust management of blockchain to study the decentralized solution of trusted data sharing, and to solve the problem of storage of shared data, propose the bloc...
详细信息
Malware signature should be compact to reduce the possibility of being bypassed. However, previous studies combine unnecessary features because they cannot distinguish associated features. This study proposes an algor...
详细信息
ISBN:
(纸本)9781665421973
Malware signature should be compact to reduce the possibility of being bypassed. However, previous studies combine unnecessary features because they cannot distinguish associated features. This study proposes an algorithm for generating effective and compact signatures, taking into account the association of features. The proposed algorithm is based on a greedy approach that chooses a feature with the lowest association with the currently combined features. As a result of conducting a family detection experiment on 4,443 samples consisting of 869 families, it was confirmed that the signature size was about 90 times smaller and the performance was 2.8% higher than in the existing method.
Nowadays, multimedia contents like images are easily distributed globally because of the widespread usage of information and communication technologies. Digital watermarking is an approach to prevent attacks or modifi...
详细信息
In today's era, the smart grid is the carrier of the new energy technology revolution and a very critical development stage for grid intelligence. In the process of smart grid operation, maintenance and maintenanc...
详细信息
Imbalanced datasets are prevalent in real life, which have a skewed data distribution. In the research of imbalanced datasets, the inherent patterns of minority class affect the generalization performance of Machine L...
详细信息
ISBN:
(纸本)9789819756650;9789819756667
Imbalanced datasets are prevalent in real life, which have a skewed data distribution. In the research of imbalanced datasets, the inherent patterns of minority class affect the generalization performance of Machine Learning models. The re-weighting/re-samplingmethods are used to address the issue of imbalanced datasets. However, the above methods are formulated based on empirical and heuristic rules. A theoretical framework is required in the presentation of the re-weighting/re-sampling methods. Category Theory is used as a framework to analyze the training process of Machine Learning models on imbalanced data. Based on the analysis, a method named Dynamic Weight Adjustment (DWA) is proposed to improve the generalization performance of models on imbalanced data. The imbalanced datasets for binary classification and multi-classification problems are used to verify the effectiveness of the DWA. The results of the DWA are the best in both imbalanced datasets for binary classification and multi-classification problems.
In the context of the era of bigdata, it not only creates a lot of opportunities for the development of modern society but also poses more challenges. If the society wants to develop better, it must take good technic...
详细信息
IEC 61499 and bigdata systems, which are also distributed architecture, but they belong to two independent and unrelated platforms due to different network and control concepts. With the development of intelligent ma...
详细信息
With the increasing number of bigdata applications, large amounts of valuable data are distributed in different organizations or regions. Federated Learning (FL) enables collaborative model training without sharing s...
详细信息
ISBN:
(纸本)9798350381993;9798350382006
With the increasing number of bigdata applications, large amounts of valuable data are distributed in different organizations or regions. Federated Learning (FL) enables collaborative model training without sharing sensitive data and is widely used in AI medical diagnosis, economy, and autonomous driving scenarios. However, it still leaks the privacy from the gradient exchange in federated learning. What's worse, state-of-the-art work, such as Batchcrypt, still suffers from computational overhead due to a considerable amount of computation and communication costs caused by homomorphic encryption. Therefore, we propose a novel symmetric key-based homomorphic encryption scheme, Sym-Fed. To unleash the power of symmetric encryption in federated learning, we combine random masking with symmetric encryption and keep the homomorphic property during the gradient exchange in the federated learning process. Finally, the security analysis and experimental results on real workloads show that our design achieves performance improvement 6x to 668x and reduces the communication overhead 1.2x to 107x compared with the state-of-the-art work, BatchCrypt and FATE, without model accuracy degradation and security compromise.
Presto is one of the most popular OLAP (Online Analytical Processing) query engines. A Presto cluster consists of multiple worker nodes that execute tasks. As of today, Presto's workload placement is not resource-...
详细信息
ISBN:
(纸本)9781665421973
Presto is one of the most popular OLAP (Online Analytical Processing) query engines. A Presto cluster consists of multiple worker nodes that execute tasks. As of today, Presto's workload placement is not resource-aware, which leads to wildly unbalanced resource usages across workers, jeopardizing business-important metrics such as reliability, performance and efficiency. We aim to add intelligence to Presto workload placement by making it resource-aware. Referred to as Presto Pack, this feature is mainly implemented on Presto coordinator. It takes into account of two resource types: memory and CPU on worker nodes. Evaluation results show that Presto Pack can achieve much more balanced CPU and memory usage across workers, less reliability issues, and higher throughput. Specifically, Presto Pack can improve overall query throughput by 23% on a per-server basis, leading to higher levels of concurrency and reliability.
In the era of artificial intelligence, how recommendation algorithms can accurately and effectively solve the problem of matching individuals' interests and needs in the data ocean, so that they can quickly obtain...
详细信息
暂无评论