Eureka: Edge-Based Discovery of Training Data for Machine Learning

The generation of high-quality training data has become the key bottleneck in the use of deep learning across many domains. In this paper, we describe Eureka, an interactive system that leverages edge computing and early discard to greatly improve the productivity of experts in the construction of a...

Full description

Saved in:
Bibliographic Details
Published in:IEEE internet computing 2019-07, Vol.23 (4), p.35-42
Main Authors: Feng, Ziqiang, George, Shilpa, Harkes, Jan, Klatzky, Roberta L., Satyanarayanan, Mahadev, Pillai, Padmanabhan
Format: Article
Language:English
Subjects:
Bandwidth
Cloud computing
Graphical user interfaces
Support vector machines
Training
Training data
Citations: Items that this one cites
Items that cite this one
Online Access:Request full text
Tags: Add Tag
No Tags, Be the first to tag this record!
cited_by cdi_FETCH-LOGICAL-c263t-d664ba63537012ebe0d9549fb89d4f35908c971ca076d59f22c4e9de1fc03d203
cites cdi_FETCH-LOGICAL-c263t-d664ba63537012ebe0d9549fb89d4f35908c971ca076d59f22c4e9de1fc03d203
container_end_page 42
container_issue 4
container_start_page 35
container_title IEEE internet computing
container_volume 23
creator Feng, Ziqiang
George, Shilpa
Harkes, Jan
Klatzky, Roberta L.
Satyanarayanan, Mahadev
Pillai, Padmanabhan
description The generation of high-quality training data has become the key bottleneck in the use of deep learning across many domains. In this paper, we describe Eureka, an interactive system that leverages edge computing and early discard to greatly improve the productivity of experts in the construction of a labeled dataset. Our experimental results show that Eureka reduces the labeling effort needed to construct a training set by two orders of magnitude relative to a brute-force approach.
doi_str_mv 10.1109/MIC.2019.2892941
format article
fullrecord <record><control><sourceid>crossref_CHZPO</sourceid><recordid>TN_cdi_crossref_primary_10_1109_MIC_2019_2892941</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><ieee_id>8611150</ieee_id><sourcerecordid>10_1109_MIC_2019_2892941</sourcerecordid><originalsourceid>FETCH-LOGICAL-c263t-d664ba63537012ebe0d9549fb89d4f35908c971ca076d59f22c4e9de1fc03d203</originalsourceid><addsrcrecordid>eNo9kEtLAzEUhYMoWKt7wU3-QMZ7k3kk7uy0aqHFTV0PaXJTx8eMJFXov--UFlfnwHksPsZuETJEMPfLeZ1JQJNJbaTJ8YyNcBABqPB88KCNqDTgJbtK6QMAtJY4YpPZb6RP-8BnfkNiYhN5Pm2T6_8o7ngf-Cratmu7DZ_areWhj3xp3XvbEV-QjYfkml0E-5Xo5qRj9vY0W9UvYvH6PK8fF8LJUm2FL8t8bUtVqApQ0prAmyI3Ya2Nz4MqDGhnKnQWqtIXJkjpcjKeMDhQXoIaMzj-utinFCk0P7H9tnHXIDQHBs3AoDkwaE4MhsndcdIS0X9dl4hYgNoDQCVWfA</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype></control><display><type>article</type><title>Eureka: Edge-Based Discovery of Training Data for Machine Learning</title><source>IEEE Xplore All Conference Series</source><creator>Feng, Ziqiang ; George, Shilpa ; Harkes, Jan ; Klatzky, Roberta L. ; Satyanarayanan, Mahadev ; Pillai, Padmanabhan</creator><creatorcontrib>Feng, Ziqiang ; George, Shilpa ; Harkes, Jan ; Klatzky, Roberta L. ; Satyanarayanan, Mahadev ; Pillai, Padmanabhan</creatorcontrib><description>The generation of high-quality training data has become the key bottleneck in the use of deep learning across many domains. In this paper, we describe Eureka, an interactive system that leverages edge computing and early discard to greatly improve the productivity of experts in the construction of a labeled dataset. Our experimental results show that Eureka reduces the labeling effort needed to construct a training set by two orders of magnitude relative to a brute-force approach.</description><identifier>ISSN: 1089-7801</identifier><identifier>EISSN: 1941-0131</identifier><identifier>DOI: 10.1109/MIC.2019.2892941</identifier><identifier>CODEN: IICOFX</identifier><language>eng</language><publisher>IEEE</publisher><subject>Bandwidth ; Cloud computing ; Graphical user interfaces ; Support vector machines ; Training ; Training data</subject><ispartof>IEEE internet computing, 2019-07, Vol.23 (4), p.35-42</ispartof><lds50>peer_reviewed</lds50><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed><citedby>FETCH-LOGICAL-c263t-d664ba63537012ebe0d9549fb89d4f35908c971ca076d59f22c4e9de1fc03d203</citedby><cites>FETCH-LOGICAL-c263t-d664ba63537012ebe0d9549fb89d4f35908c971ca076d59f22c4e9de1fc03d203</cites></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><linktohtml>$$Uhttps://ieeexplore.ieee.org/document/8611150$$EHTML$$P50$$Gieee$$H</linktohtml><link.rule.ids>314,780,784,27924,27925,54555,54796,54932</link.rule.ids><linktorsrc>$$Uhttps://ieeexplore.ieee.org/document/8611150$$EView_record_in_IEEE$$FView_record_in_$$GIEEE</linktorsrc></links><search><creatorcontrib>Feng, Ziqiang</creatorcontrib><creatorcontrib>George, Shilpa</creatorcontrib><creatorcontrib>Harkes, Jan</creatorcontrib><creatorcontrib>Klatzky, Roberta L.</creatorcontrib><creatorcontrib>Satyanarayanan, Mahadev</creatorcontrib><creatorcontrib>Pillai, Padmanabhan</creatorcontrib><title>Eureka: Edge-Based Discovery of Training Data for Machine Learning</title><title>IEEE internet computing</title><addtitle>MIC</addtitle><description>The generation of high-quality training data has become the key bottleneck in the use of deep learning across many domains. In this paper, we describe Eureka, an interactive system that leverages edge computing and early discard to greatly improve the productivity of experts in the construction of a labeled dataset. Our experimental results show that Eureka reduces the labeling effort needed to construct a training set by two orders of magnitude relative to a brute-force approach.</description><subject>Bandwidth</subject><subject>Cloud computing</subject><subject>Graphical user interfaces</subject><subject>Support vector machines</subject><subject>Training</subject><subject>Training data</subject><issn>1089-7801</issn><issn>1941-0131</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2019</creationdate><recordtype>article</recordtype><recordid>eNo9kEtLAzEUhYMoWKt7wU3-QMZ7k3kk7uy0aqHFTV0PaXJTx8eMJFXov--UFlfnwHksPsZuETJEMPfLeZ1JQJNJbaTJ8YyNcBABqPB88KCNqDTgJbtK6QMAtJY4YpPZb6RP-8BnfkNiYhN5Pm2T6_8o7ngf-Cratmu7DZ_areWhj3xp3XvbEV-QjYfkml0E-5Xo5qRj9vY0W9UvYvH6PK8fF8LJUm2FL8t8bUtVqApQ0prAmyI3Ya2Nz4MqDGhnKnQWqtIXJkjpcjKeMDhQXoIaMzj-utinFCk0P7H9tnHXIDQHBs3AoDkwaE4MhsndcdIS0X9dl4hYgNoDQCVWfA</recordid><startdate>201907</startdate><enddate>201907</enddate><creator>Feng, Ziqiang</creator><creator>George, Shilpa</creator><creator>Harkes, Jan</creator><creator>Klatzky, Roberta L.</creator><creator>Satyanarayanan, Mahadev</creator><creator>Pillai, Padmanabhan</creator><general>IEEE</general><scope>97E</scope><scope>RIA</scope><scope>RIE</scope><scope>AAYXX</scope><scope>CITATION</scope></search><sort><creationdate>201907</creationdate><title>Eureka: Edge-Based Discovery of Training Data for Machine Learning</title><author>Feng, Ziqiang ; George, Shilpa ; Harkes, Jan ; Klatzky, Roberta L. ; Satyanarayanan, Mahadev ; Pillai, Padmanabhan</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-LOGICAL-c263t-d664ba63537012ebe0d9549fb89d4f35908c971ca076d59f22c4e9de1fc03d203</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2019</creationdate><topic>Bandwidth</topic><topic>Cloud computing</topic><topic>Graphical user interfaces</topic><topic>Support vector machines</topic><topic>Training</topic><topic>Training data</topic><toplevel>peer_reviewed</toplevel><toplevel>online_resources</toplevel><creatorcontrib>Feng, Ziqiang</creatorcontrib><creatorcontrib>George, Shilpa</creatorcontrib><creatorcontrib>Harkes, Jan</creatorcontrib><creatorcontrib>Klatzky, Roberta L.</creatorcontrib><creatorcontrib>Satyanarayanan, Mahadev</creatorcontrib><creatorcontrib>Pillai, Padmanabhan</creatorcontrib><collection>IEEE All-Society Periodicals Package (ASPP) 2005–Present</collection><collection>IEEE All-Society Periodicals Package (ASPP) 1998–Present</collection><collection>IEEE Electronic Library Online</collection><collection>CrossRef</collection><jtitle>IEEE internet computing</jtitle></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext_linktorsrc</fulltext></delivery><addata><au>Feng, Ziqiang</au><au>George, Shilpa</au><au>Harkes, Jan</au><au>Klatzky, Roberta L.</au><au>Satyanarayanan, Mahadev</au><au>Pillai, Padmanabhan</au><format>journal</format><genre>article</genre><ristype>JOUR</ristype><atitle>Eureka: Edge-Based Discovery of Training Data for Machine Learning</atitle><jtitle>IEEE internet computing</jtitle><stitle>MIC</stitle><date>2019-07</date><risdate>2019</risdate><volume>23</volume><issue>4</issue><spage>35</spage><epage>42</epage><pages>35-42</pages><issn>1089-7801</issn><eissn>1941-0131</eissn><coden>IICOFX</coden><abstract>The generation of high-quality training data has become the key bottleneck in the use of deep learning across many domains. In this paper, we describe Eureka, an interactive system that leverages edge computing and early discard to greatly improve the productivity of experts in the construction of a labeled dataset. Our experimental results show that Eureka reduces the labeling effort needed to construct a training set by two orders of magnitude relative to a brute-force approach.</abstract><pub>IEEE</pub><doi>10.1109/MIC.2019.2892941</doi><tpages>8</tpages><oa>free_for_read</oa></addata></record>
fulltext fulltext_linktorsrc
identifier ISSN: 1089-7801
ispartof IEEE internet computing, 2019-07, Vol.23 (4), p.35-42
issn 1089-7801
1941-0131
language eng
recordid cdi_crossref_primary_10_1109_MIC_2019_2892941
source IEEE Xplore All Conference Series
subjects Bandwidth
Cloud computing
Graphical user interfaces
Support vector machines
Training
Training data
title Eureka: Edge-Based Discovery of Training Data for Machine Learning
url http://sfxeu10.hosted.exlibrisgroup.com/loughborough?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2024-12-24T18%3A47%3A43IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-crossref_CHZPO&rft_val_fmt=info:ofi/fmt:kev:mtx:journal&rft.genre=article&rft.atitle=Eureka:%20Edge-Based%20Discovery%20of%20Training%20Data%20for%20Machine%20Learning&rft.jtitle=IEEE%20internet%20computing&rft.au=Feng,%20Ziqiang&rft.date=2019-07&rft.volume=23&rft.issue=4&rft.spage=35&rft.epage=42&rft.pages=35-42&rft.issn=1089-7801&rft.eissn=1941-0131&rft.coden=IICOFX&rft_id=info:doi/10.1109/MIC.2019.2892941&rft_dat=%3Ccrossref_CHZPO%3E10_1109_MIC_2019_2892941%3C/crossref_CHZPO%3E%3Cgrp_id%3Ecdi_FETCH-LOGICAL-c263t-d664ba63537012ebe0d9549fb89d4f35908c971ca076d59f22c4e9de1fc03d203%3C/grp_id%3E%3Coa%3E%3C/oa%3E%3Curl%3E%3C/url%3E&rft_id=info:oai/&rft_id=info:pmid/&rft_ieee_id=8611150&rfr_iscdi=true