Using Time-Series Privileged Information for Provably Efficient Learning of Prediction Models
We study prediction of future outcomes with supervised models that use privileged information during learning. The privileged information comprises samples of time series observed between the baseline time of prediction and the future outcome; this information is only available at training time whic...
Gespeichert in:
Veröffentlicht in: | arXiv.org 2022-05 |
---|---|
Hauptverfasser: | , , , , , |
Format: | Artikel |
Sprache: | eng |
Schlagworte: | |
Online-Zugang: | Volltext |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
container_end_page | |
---|---|
container_issue | |
container_start_page | |
container_title | arXiv.org |
container_volume | |
creator | Karlsson, Rickard K A Willbo, Martin Hussain, Zeshan Krishnan, Rahul G Sontag, David Johansson, Fredrik D |
description | We study prediction of future outcomes with supervised models that use privileged information during learning. The privileged information comprises samples of time series observed between the baseline time of prediction and the future outcome; this information is only available at training time which differs from the traditional supervised learning. Our question is when using this privileged data leads to more sample-efficient learning of models that use only baseline data for predictions at test time. We give an algorithm for this setting and prove that when the time series are drawn from a non-stationary Gaussian-linear dynamical system of fixed horizon, learning with privileged information is more efficient than learning without it. On synthetic data, we test the limits of our algorithm and theory, both when our assumptions hold and when they are violated. On three diverse real-world datasets, we show that our approach is generally preferable to classical learning, particularly when data is scarce. Finally, we relate our estimator to a distillation approach both theoretically and empirically. |
format | Article |
fullrecord | <record><control><sourceid>proquest</sourceid><recordid>TN_cdi_proquest_journals_2588154446</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2588154446</sourcerecordid><originalsourceid>FETCH-proquest_journals_25881544463</originalsourceid><addsrcrecordid>eNqNjNEKgjAYRkcQJOU7DLoWdFPzPoyCgiC7DDH9J7_MrTYVevtW9ABdfQfO4ZsRj3EeBVnM2IL41nZhGLJ0w5KEe-R2tahaWmAPwQUMgqVngxNKaKGhByW06asBtaKOnNJTdZcvmguBNYIa6BEqoz4XWjgNDdbf-qQbkHZF5qKSFvzfLsl6lxfbffAw-jmCHcpOj0Y5VbIky6IkjuOU_1e9AQ7pQ9s</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>2588154446</pqid></control><display><type>article</type><title>Using Time-Series Privileged Information for Provably Efficient Learning of Prediction Models</title><source>Free E- Journals</source><creator>Karlsson, Rickard K A ; Willbo, Martin ; Hussain, Zeshan ; Krishnan, Rahul G ; Sontag, David ; Johansson, Fredrik D</creator><creatorcontrib>Karlsson, Rickard K A ; Willbo, Martin ; Hussain, Zeshan ; Krishnan, Rahul G ; Sontag, David ; Johansson, Fredrik D</creatorcontrib><description>We study prediction of future outcomes with supervised models that use privileged information during learning. The privileged information comprises samples of time series observed between the baseline time of prediction and the future outcome; this information is only available at training time which differs from the traditional supervised learning. Our question is when using this privileged data leads to more sample-efficient learning of models that use only baseline data for predictions at test time. We give an algorithm for this setting and prove that when the time series are drawn from a non-stationary Gaussian-linear dynamical system of fixed horizon, learning with privileged information is more efficient than learning without it. On synthetic data, we test the limits of our algorithm and theory, both when our assumptions hold and when they are violated. On three diverse real-world datasets, we show that our approach is generally preferable to classical learning, particularly when data is scarce. Finally, we relate our estimator to a distillation approach both theoretically and empirically.</description><identifier>EISSN: 2331-8422</identifier><language>eng</language><publisher>Ithaca: Cornell University Library, arXiv.org</publisher><subject>Algorithms ; Distillation ; Prediction models ; Supervised learning ; Testing time ; Time series</subject><ispartof>arXiv.org, 2022-05</ispartof><rights>2022. This work is published under http://arxiv.org/licenses/nonexclusive-distrib/1.0/ (the “License”). Notwithstanding the ProQuest Terms and Conditions, you may use this content in accordance with the terms of the License.</rights><oa>free_for_read</oa><woscitedreferencessubscribed>false</woscitedreferencessubscribed></display><links><openurl>$$Topenurl_article</openurl><openurlfulltext>$$Topenurlfull_article</openurlfulltext><thumbnail>$$Tsyndetics_thumb_exl</thumbnail><link.rule.ids>776,780</link.rule.ids></links><search><creatorcontrib>Karlsson, Rickard K A</creatorcontrib><creatorcontrib>Willbo, Martin</creatorcontrib><creatorcontrib>Hussain, Zeshan</creatorcontrib><creatorcontrib>Krishnan, Rahul G</creatorcontrib><creatorcontrib>Sontag, David</creatorcontrib><creatorcontrib>Johansson, Fredrik D</creatorcontrib><title>Using Time-Series Privileged Information for Provably Efficient Learning of Prediction Models</title><title>arXiv.org</title><description>We study prediction of future outcomes with supervised models that use privileged information during learning. The privileged information comprises samples of time series observed between the baseline time of prediction and the future outcome; this information is only available at training time which differs from the traditional supervised learning. Our question is when using this privileged data leads to more sample-efficient learning of models that use only baseline data for predictions at test time. We give an algorithm for this setting and prove that when the time series are drawn from a non-stationary Gaussian-linear dynamical system of fixed horizon, learning with privileged information is more efficient than learning without it. On synthetic data, we test the limits of our algorithm and theory, both when our assumptions hold and when they are violated. On three diverse real-world datasets, we show that our approach is generally preferable to classical learning, particularly when data is scarce. Finally, we relate our estimator to a distillation approach both theoretically and empirically.</description><subject>Algorithms</subject><subject>Distillation</subject><subject>Prediction models</subject><subject>Supervised learning</subject><subject>Testing time</subject><subject>Time series</subject><issn>2331-8422</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2022</creationdate><recordtype>article</recordtype><sourceid>BENPR</sourceid><recordid>eNqNjNEKgjAYRkcQJOU7DLoWdFPzPoyCgiC7DDH9J7_MrTYVevtW9ABdfQfO4ZsRj3EeBVnM2IL41nZhGLJ0w5KEe-R2tahaWmAPwQUMgqVngxNKaKGhByW06asBtaKOnNJTdZcvmguBNYIa6BEqoz4XWjgNDdbf-qQbkHZF5qKSFvzfLsl6lxfbffAw-jmCHcpOj0Y5VbIky6IkjuOU_1e9AQ7pQ9s</recordid><startdate>20220505</startdate><enddate>20220505</enddate><creator>Karlsson, Rickard K A</creator><creator>Willbo, Martin</creator><creator>Hussain, Zeshan</creator><creator>Krishnan, Rahul G</creator><creator>Sontag, David</creator><creator>Johansson, Fredrik D</creator><general>Cornell University Library, arXiv.org</general><scope>8FE</scope><scope>8FG</scope><scope>ABJCF</scope><scope>ABUWG</scope><scope>AFKRA</scope><scope>AZQEC</scope><scope>BENPR</scope><scope>BGLVJ</scope><scope>CCPQU</scope><scope>DWQXO</scope><scope>HCIFZ</scope><scope>L6V</scope><scope>M7S</scope><scope>PIMPY</scope><scope>PQEST</scope><scope>PQQKQ</scope><scope>PQUKI</scope><scope>PRINS</scope><scope>PTHSS</scope></search><sort><creationdate>20220505</creationdate><title>Using Time-Series Privileged Information for Provably Efficient Learning of Prediction Models</title><author>Karlsson, Rickard K A ; Willbo, Martin ; Hussain, Zeshan ; Krishnan, Rahul G ; Sontag, David ; Johansson, Fredrik D</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-proquest_journals_25881544463</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2022</creationdate><topic>Algorithms</topic><topic>Distillation</topic><topic>Prediction models</topic><topic>Supervised learning</topic><topic>Testing time</topic><topic>Time series</topic><toplevel>online_resources</toplevel><creatorcontrib>Karlsson, Rickard K A</creatorcontrib><creatorcontrib>Willbo, Martin</creatorcontrib><creatorcontrib>Hussain, Zeshan</creatorcontrib><creatorcontrib>Krishnan, Rahul G</creatorcontrib><creatorcontrib>Sontag, David</creatorcontrib><creatorcontrib>Johansson, Fredrik D</creatorcontrib><collection>ProQuest SciTech Collection</collection><collection>ProQuest Technology Collection</collection><collection>Materials Science & Engineering Collection</collection><collection>ProQuest Central (Alumni Edition)</collection><collection>ProQuest Central UK/Ireland</collection><collection>ProQuest Central Essentials</collection><collection>ProQuest Central</collection><collection>Technology Collection</collection><collection>ProQuest One Community College</collection><collection>ProQuest Central Korea</collection><collection>SciTech Premium Collection</collection><collection>ProQuest Engineering Collection</collection><collection>Engineering Database</collection><collection>Publicly Available Content Database</collection><collection>ProQuest One Academic Eastern Edition (DO NOT USE)</collection><collection>ProQuest One Academic</collection><collection>ProQuest One Academic UKI Edition</collection><collection>ProQuest Central China</collection><collection>Engineering Collection</collection></facets><delivery><delcategory>Remote Search Resource</delcategory><fulltext>fulltext</fulltext></delivery><addata><au>Karlsson, Rickard K A</au><au>Willbo, Martin</au><au>Hussain, Zeshan</au><au>Krishnan, Rahul G</au><au>Sontag, David</au><au>Johansson, Fredrik D</au><format>book</format><genre>document</genre><ristype>GEN</ristype><atitle>Using Time-Series Privileged Information for Provably Efficient Learning of Prediction Models</atitle><jtitle>arXiv.org</jtitle><date>2022-05-05</date><risdate>2022</risdate><eissn>2331-8422</eissn><abstract>We study prediction of future outcomes with supervised models that use privileged information during learning. The privileged information comprises samples of time series observed between the baseline time of prediction and the future outcome; this information is only available at training time which differs from the traditional supervised learning. Our question is when using this privileged data leads to more sample-efficient learning of models that use only baseline data for predictions at test time. We give an algorithm for this setting and prove that when the time series are drawn from a non-stationary Gaussian-linear dynamical system of fixed horizon, learning with privileged information is more efficient than learning without it. On synthetic data, we test the limits of our algorithm and theory, both when our assumptions hold and when they are violated. On three diverse real-world datasets, we show that our approach is generally preferable to classical learning, particularly when data is scarce. Finally, we relate our estimator to a distillation approach both theoretically and empirically.</abstract><cop>Ithaca</cop><pub>Cornell University Library, arXiv.org</pub><oa>free_for_read</oa></addata></record> |
fulltext | fulltext |
identifier | EISSN: 2331-8422 |
ispartof | arXiv.org, 2022-05 |
issn | 2331-8422 |
language | eng |
recordid | cdi_proquest_journals_2588154446 |
source | Free E- Journals |
subjects | Algorithms Distillation Prediction models Supervised learning Testing time Time series |
title | Using Time-Series Privileged Information for Provably Efficient Learning of Prediction Models |
url | https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-02-13T01%3A49%3A06IST&url_ver=Z39.88-2004&url_ctx_fmt=infofi/fmt:kev:mtx:ctx&rfr_id=info:sid/primo.exlibrisgroup.com:primo3-Article-proquest&rft_val_fmt=info:ofi/fmt:kev:mtx:book&rft.genre=document&rft.atitle=Using%20Time-Series%20Privileged%20Information%20for%20Provably%20Efficient%20Learning%20of%20Prediction%20Models&rft.jtitle=arXiv.org&rft.au=Karlsson,%20Rickard%20K%20A&rft.date=2022-05-05&rft.eissn=2331-8422&rft_id=info:doi/&rft_dat=%3Cproquest%3E2588154446%3C/proquest%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=2588154446&rft_id=info:pmid/&rfr_iscdi=true |