Dynamic Adaptive Replacement Policy in Shared Last-Level Cache of DRAM/PCM Hybrid Memory for Big Data Storage

Jia, Gangyong; Han, Guangjie; Jiang, Jinfang; Liu, Li

doi:10.1109/tii.2016.2645941

Cited by 55 publications

(22 citation statements)

References 20 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…. , A y (y kinds of learning algorithms) Output: A (detected noisy subset of E) (1) form n disjoint almost equally sized subset of E i , where (6) induce H j based on examples in E t and A j (7) end for (8) for every e ∈ E i do (9) ErrorCounter ← 0 (10)…”

Section: Algorithm 1 Majority Filtering Algorithmmentioning

confidence: 99%

“…. y do (7) induce H pj based on examples in E t and A j (8) end for (9) for every e ∈ E pi do (10) ErrorCounter ← 0 (11) for j = 1,. . .…”

Section: Majorityfiltering_majorityfiltering (Mf Mf ) Input: E (Trainmentioning

confidence: 99%

“…For example, sensor involved applications (such as WSN and IoT) may make noises due to the intrinsic instability of sensors [6,7]. In addition, big data further contribute to the emergence of noise [8]. When training data are noisy, the performance of learning based on it will be degraded.…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Enhanced Label Noise Filtering with Multiple Voting

et al. 2019

View full text Add to dashboard Cite

Label noises exist in many applications, and their presence can degrade learning performance. Researchers usually use filters to identify and eliminate them prior to training. The ensemble learning based filter (EnFilter) is the most widely used filter. According to the voting mechanism, EnFilter is mainly divided into two types: single-voting based (SVFilter) and multiple-voting based (MVFilter). In general, MVFilter is more often preferred because multiple-voting could address the intrinsic limitations of single-voting. However, the most important unsolved issue in MVFilter is how to determine the optimal decision point (ODP). Conceptually, the decision point is a threshold value, which determines the noise detection performance. To maximize the performance of MVFilter, we propose a novel approach to compute the optimal decision point. Our approach is data driven and cost sensitive, which determines the ODP based on the given noisy training dataset and noise misrecognition cost matrix. The core idea of our approach is to estimate the mislabeled data probability distributions, based on which the expected cost of each possible decision point could be inferred. Experimental results on a set of benchmark datasets illustrate the utility of our proposed approach.

show abstract

Section: Algorithm 1 Majority Filtering Algorithmmentioning

confidence: 99%

“…. y do (7) induce H pj based on examples in E t and A j (8) end for (9) for every e ∈ E pi do (10) ErrorCounter ← 0 (11) for j = 1,. . .…”

Section: Majorityfiltering_majorityfiltering (Mf Mf ) Input: E (Trainmentioning

confidence: 99%

See 1 more Smart Citation

Enhanced Label Noise Filtering with Multiple Voting

et al. 2019

View full text Add to dashboard Cite

show abstract

“…Anyone can check the validity of a ring signature when , , and 1 to are given. All the members in the ring can sign a message with their own private keys and the others' public keys without the agreement of them [17][18][19][20][21][22][23]. With ring signature, the message can be transmitted securely and the identity of the actual signer is protected as well.…”

Section: Preliminariesmentioning

confidence: 99%

Cloud Based Data Protection in Anonymously Controlled SDN

Shen

Lai

et al. 2018

Security and Communication Networks

View full text Add to dashboard Cite

Nowadays, Software Defined Network (SDN) develops rapidly for its novel structure which separates the control plane and the data plane of network devices. Many researchers devoted themselves to the study of such a special network. However, some limitations restrict the development of SDN. On the one hand, the single controller in the conventional model bears all threats, and the corruption of it will result in network paralysis. On the other hand, the data will be increasing more in SDN switches in the data plane, while the storage space of these switches is limited. In order to solve the mentioned issues, we propose two corresponding protocols in this paper. Specifically, one is an anonymous protocol in the control plane, and the other is a verifiable outsourcing protocol in the data plane. The evaluation indicates that our protocol is correct, secure, and efficient.

show abstract

“…For example, there are proposals that prioritize geographic area coverage as well as barrier coverage, rather than the coverage of discrete targets, and allow for network nodes to move after deployment [11], In [12], a maximum covers algorithm using mixed integer programming (MC-MIP) was proposed, Based on the output of MC-MIP, nodes are organized into disjoint set covers (DSC) which are activated successively, In [13], a greedy algorithm designed for maximum set covers (MSCGreedy) was proposed, Cover sets generated from MSCGreedy are not required to be disjoint and are allowed to operate in different time intervals, Compared with MC-MIP, MSC-Greedy produces better results in terms of network lifetime, since the solution space of DSC problem is included in the solution space of the maximum set cover (MSC) problem, However, current researches on sensor nodes deployment mainly focus on ensuring sensing areas connectivity through optimizing sensing coverage probability (CP), For example, [14] devised the nodes deployed model based on Gaussian distribution and concluded that sensing CP was improved with the increase of standard deviation of the deployed errors; [15] introduced missing probability to obtain optimal number of deployed sensor nodes and deployment according to the terrain model and the deployed precision, and reformed the algorithmproposed in [16], which optimized CP to improve deployed efficiency in terms of uncertainty of sensing probability.…”

Section: Related Workmentioning

confidence: 99%

Research on Wireless Sensor Actuator Distribution Algorithm Meeting Maximum Request Expectation

Jiang

Zhang

et al. 2017

Int. J. Onl. Eng.

View full text Add to dashboard Cite

Abstract-In wireless sensor actuator networks, actuator nodes are very important to network coverage for wireless sensor, So according to the characteristics that wireless communication is easy to be affected by the environment, in this wireless sensor actuator network the actuator nodes are abstracted to be service providers and the sensor nodes are abstracted to be service consumers, Furthermore, the messages transferred from actuator nodes to sensor nodes are abstracted to be service requests and the coverage of sensor nodes by actuator nodes is service providing, Moreover, the service faulty ratio of actuator nodes is defined in this paper, Based on that, an actuator node coverage algorithm that satisfies the maximum request expectation is put forwarded, Its goal is to solve the mathematical expectation of maximizing the number of service requests with fixed number of sensor nodes and actuator nodes, So the optimal position of actuator nodes could be identified, The theoretic proof and simulation result show that this algorithm is more efficient compared with similar algorithms.

show abstract

Dynamic Adaptive Replacement Policy in Shared Last-Level Cache of DRAM/PCM Hybrid Memory for Big Data Storage

Cited by 55 publications

References 20 publications

Enhanced Label Noise Filtering with Multiple Voting

Enhanced Label Noise Filtering with Multiple Voting

Cloud Based Data Protection in Anonymously Controlled SDN

Research on Wireless Sensor Actuator Distribution Algorithm Meeting Maximum Request Expectation

Contact Info

Product

Resources

About