國立虎尾科技大學 |

Cyberinfrastructure Knowledge Network : = Optimizing Edge-Cloud-Continuum for Efficient Data Capture, Modeling, and Deployment.

Record Type:	Language materials, manuscript : Monograph/item
Title/Author:	Cyberinfrastructure Knowledge Network :/
Reminder of title:	Optimizing Edge-Cloud-Continuum for Efficient Data Capture, Modeling, and Deployment.
Author:	Withana, Sachith Danushka.
Description:	1 online resource (134 pages)
Notes:	Source: Dissertations Abstracts International, Volume: 85-12, Section: B.
Contained By:	Dissertations Abstracts International85-12B.
Subject:	Computer engineering. -
Online resource:	click for full text (PQDT)
ISBN:	9798383056738

Cyberinfrastructure Knowledge Network : = Optimizing Edge-Cloud-Continuum for Efficient Data Capture, Modeling, and Deployment.
Withana, Sachith Danushka.

Cyberinfrastructure Knowledge Network :Optimizing Edge-Cloud-Continuum for Efficient Data Capture, Modeling, and Deployment. - 1 online resource (134 pages)

Source: Dissertations Abstracts International, Volume: 85-12, Section: B.

Thesis (Ph.D.)--Indiana University, 2024.

Includes bibliographical references

The edge-cloud continuum introduces new challenges to distributed computing. It incorporates a varied landscape of computing resources, from edge devices to centralized cloud data centers. Our approach is to strategically distribute AI-driven computing tasks along the continuum, considering factors such as latency, data sensitivity, user quality of experience, and computational intensity. At one end of the spectrum, edge devices, such as IoT sensors and smartphones, capture and process data in real-time with low latency. At the other end, cloud data centers provide extensive computational power and storage capacity. In between are edge computing nodes, which offer a balance between proximity to edge devices and computational capabilities.This dissertation presents a framework, named Cyberinfrastructure Knowledge Network(CKN), which efficiently distributes computing across the edge-cloud continuum. CKN does this by representing historical data and provenance information that it then uses to optimize the continuum. By processing data streams efficiently, data can be aggregated, and alerts can be generated in real-time before being inserted into the CKN knowledge graph for post-analysis. The framework uses animal ecology camera traps for animal identification as a use case and demonstrates the system's ability to leverage historical data for efficient model placement at the edge. We show that the CKN framework can detect and respond rapidly when a new AI inference model is needed.This dissertation advances a semi-automated model card solution for machine learning pipelines in scientific problem domains. The solution, called Patra Model Cards, captures and stores data flow from model creation to deployment and completes a structured model card in a semi-automated manner. The objective is the deployment of accountable, fair, and explainable AI models at the edge while leveraging historical information on how these models performed in various contexts.

Electronic reproduction.
Ann Arbor, Mich. :
ProQuest,
2024

Mode of access: World Wide Web

ISBN: 9798383056738Subjects--Topical Terms:

569006
Computer engineering.
Subjects--Index Terms:

Edge computingIndex Terms--Genre/Form:

554714
Electronic books.

Cyberinfrastructure Knowledge Network : = Optimizing Edge-Cloud-Continuum for Efficient Data Capture, Modeling, and Deployment.
LDR:03421ntm a22003977 4500 001 1150205
005 20241022111614.5
006 m o d
007 cr bn ---uuuuu
008 250605s2024 xx obm 000 0 eng d
020 $a 9798383056738
035 $a (MiAaPQ)AAI31298170
035 $a AAI31298170
040 $a MiAaPQ $b eng $c MiAaPQ $d NTU
100 1 $a Withana, Sachith Danushka. $3 1476645
245 1 0 $a Cyberinfrastructure Knowledge Network : $b Optimizing Edge-Cloud-Continuum for Efficient Data Capture, Modeling, and Deployment.
264 0 $c 2024
300 $a 1 online resource (134 pages)
336 $a text $b txt $2 rdacontent
337 $a computer $b c $2 rdamedia
338 $a online resource $b cr $2 rdacarrier
500 $a Source: Dissertations Abstracts International, Volume: 85-12, Section: B.
500 $a Advisor: Plale, Beth.
502 $a Thesis (Ph.D.)--Indiana University, 2024.
504 $a Includes bibliographical references
520 $a The edge-cloud continuum introduces new challenges to distributed computing. It incorporates a varied landscape of computing resources, from edge devices to centralized cloud data centers. Our approach is to strategically distribute AI-driven computing tasks along the continuum, considering factors such as latency, data sensitivity, user quality of experience, and computational intensity. At one end of the spectrum, edge devices, such as IoT sensors and smartphones, capture and process data in real-time with low latency. At the other end, cloud data centers provide extensive computational power and storage capacity. In between are edge computing nodes, which offer a balance between proximity to edge devices and computational capabilities.This dissertation presents a framework, named Cyberinfrastructure Knowledge Network(CKN), which efficiently distributes computing across the edge-cloud continuum. CKN does this by representing historical data and provenance information that it then uses to optimize the continuum. By processing data streams efficiently, data can be aggregated, and alerts can be generated in real-time before being inserted into the CKN knowledge graph for post-analysis. The framework uses animal ecology camera traps for animal identification as a use case and demonstrates the system's ability to leverage historical data for efficient model placement at the edge. We show that the CKN framework can detect and respond rapidly when a new AI inference model is needed.This dissertation advances a semi-automated model card solution for machine learning pipelines in scientific problem domains. The solution, called Patra Model Cards, captures and stores data flow from model creation to deployment and completes a structured model card in a semi-automated manner. The objective is the deployment of accountable, fair, and explainable AI models at the edge while leveraging historical information on how these models performed in various contexts.
533 $a Electronic reproduction. $b Ann Arbor, Mich. : $c ProQuest, $d 2024
538 $a Mode of access: World Wide Web
650 4 $a Computer engineering. $3 569006
650 4 $a Computer science. $3 573171
653 $a Edge computing
653 $a Edge-cloud continuum
653 $a Cyberinfrastructure Knowledge Network
653 $a Patra Model Cards
653 $a Data sensitivity
655 7 $a Electronic books. $2 local $3 554714
690 $a 0464
690 $a 0800
690 $a 0984
710 2 $a ProQuest Information and Learning Co. $3 1178819
710 2 $a Indiana University. $b Intelligent Systems Engineering. $3 1470189
773 0 $t Dissertations Abstracts International $g 85-12B.
856 4 0 $u http://pqdd.sinica.edu.tw/twdaoapp/servlet/advanced?query=31298170 $z click for full text (PQDT)