Paused Agent Replay Refresh

Paused Agent Replay Refresh

Reinforcement learning algorithms have become more complex since the invention of target networks. Unfortunately, target networks have not kept up with this increased complexity, instead requiring approximate solutions to be computationally feasible. These approximations increase noise in the Q-valu...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:arXiv.org 2022-09
1. Verfasser: Parr, Benjamin
Format: Artikel
Sprache:eng
Schlagworte:
Algorithms
Approximation
Complexity
Computer architecture
Machine learning
Networks
Sampling
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 Parr, Benjamin
description Reinforcement learning algorithms have become more complex since the invention of target networks. Unfortunately, target networks have not kept up with this increased complexity, instead requiring approximate solutions to be computationally feasible. These approximations increase noise in the Q-value targets and in the replay sampling distribution. Paused Agent Replay Refresh (PARR) is a drop-in replacement for target networks that supports more complex learning algorithms without this need for approximation. Using a basic Q-network architecture, and refreshing the novelty values, target values, and replay sampling distribution, PARR gets 2500 points in Montezuma's Revenge after only 30.9 million Atari frames. Finally, interpreting PARR in the context of carbon-based learning offers a new reason for sleep.
format Article
fullrecord <record><control><sourceid>proquest</sourceid><recordid>TN_cdi_proquest_journals_2718739415</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>2718739415</sourcerecordid><originalsourceid>FETCH-proquest_journals_27187394153</originalsourceid><addsrcrecordid>eNpjYuA0MjY21LUwMTLiYOAtLs4yMDAwMjM3MjU15mSQDkgsLU5NUXBMT80rUQhKLchJrARSaUWpxRk8DKxpiTnFqbxQmptB2c01xNlDt6Aov7A0tbgkPiu_tCgPKBVvZG5oYW5saWJoakycKgDRnyrr</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>2718739415</pqid></control><display><type>article</type><title>Paused Agent Replay Refresh</title><source>Freely Accessible Journals</source><creator>Parr, Benjamin</creator><creatorcontrib>Parr, Benjamin</creatorcontrib><description>Reinforcement learning algorithms have become more complex since the invention of target networks. Unfortunately, target networks have not kept up with this increased complexity, instead requiring approximate solutions to be computationally feasible. These approximations increase noise in the Q-value targets and in the replay sampling distribution. Paused Agent Replay Refresh (PARR) is a drop-in replacement for target networks that supports more complex learning algorithms without this need for approximation. Using a basic Q-network architecture, and refreshing the novelty values, target values, and replay sampling distribution, PARR gets 2500 points in Montezuma's Revenge after only 30.9 million Atari frames. Finally, interpreting PARR in the context of carbon-based learning offers a new reason for sleep.</description><identifier>EISSN: 2331-8422</identifier><language>eng</language><publisher>Ithaca: Cornell University Library, arXiv.org</publisher><subject>Algorithms ; Approximation ; Complexity ; Computer architecture ; Machine learning ; Networks ; Sampling</subject><ispartof>arXiv.org, 2022-09</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>780,784</link.rule.ids></links><search><creatorcontrib>Parr, Benjamin</creatorcontrib><title>Paused Agent Replay Refresh</title><title>arXiv.org</title><description>Reinforcement learning algorithms have become more complex since the invention of target networks. Unfortunately, target networks have not kept up with this increased complexity, instead requiring approximate solutions to be computationally feasible. These approximations increase noise in the Q-value targets and in the replay sampling distribution. Paused Agent Replay Refresh (PARR) is a drop-in replacement for target networks that supports more complex learning algorithms without this need for approximation. Using a basic Q-network architecture, and refreshing the novelty values, target values, and replay sampling distribution, PARR gets 2500 points in Montezuma's Revenge after only 30.9 million Atari frames. Finally, interpreting PARR in the context of carbon-based learning offers a new reason for sleep.</description><subject>Algorithms</subject><subject>Approximation</subject><subject>Complexity</subject><subject>Computer architecture</subject><subject>Machine learning</subject><subject>Networks</subject><subject>Sampling</subject><issn>2331-8422</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2022</creationdate><recordtype>article</recordtype><sourceid>ABUWG</sourceid><sourceid>AFKRA</sourceid><sourceid>AZQEC</sourceid><sourceid>BENPR</sourceid><sourceid>CCPQU</sourceid><sourceid>DWQXO</sourceid><recordid>eNpjYuA0MjY21LUwMTLiYOAtLs4yMDAwMjM3MjU15mSQDkgsLU5NUXBMT80rUQhKLchJrARSaUWpxRk8DKxpiTnFqbxQmptB2c01xNlDt6Aov7A0tbgkPiu_tCgPKBVvZG5oYW5saWJoakycKgDRnyrr</recordid><startdate>20220926</startdate><enddate>20220926</enddate><creator>Parr, Benjamin</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>20220926</creationdate><title>Paused Agent Replay Refresh</title><author>Parr, Benjamin</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-proquest_journals_27187394153</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2022</creationdate><topic>Algorithms</topic><topic>Approximation</topic><topic>Complexity</topic><topic>Computer architecture</topic><topic>Machine learning</topic><topic>Networks</topic><topic>Sampling</topic><toplevel>online_resources</toplevel><creatorcontrib>Parr, Benjamin</creatorcontrib><collection>ProQuest SciTech Collection</collection><collection>ProQuest Technology Collection</collection><collection>Materials Science &amp; 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>Parr, Benjamin</au><format>book</format><genre>document</genre><ristype>GEN</ristype><atitle>Paused Agent Replay Refresh</atitle><jtitle>arXiv.org</jtitle><date>2022-09-26</date><risdate>2022</risdate><eissn>2331-8422</eissn><abstract>Reinforcement learning algorithms have become more complex since the invention of target networks. Unfortunately, target networks have not kept up with this increased complexity, instead requiring approximate solutions to be computationally feasible. These approximations increase noise in the Q-value targets and in the replay sampling distribution. Paused Agent Replay Refresh (PARR) is a drop-in replacement for target networks that supports more complex learning algorithms without this need for approximation. Using a basic Q-network architecture, and refreshing the novelty values, target values, and replay sampling distribution, PARR gets 2500 points in Montezuma's Revenge after only 30.9 million Atari frames. Finally, interpreting PARR in the context of carbon-based learning offers a new reason for sleep.</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, 2022-09
issn 2331-8422
language eng
recordid cdi_proquest_journals_2718739415
source Freely Accessible Journals
subjects Algorithms
Approximation
Complexity
Computer architecture
Machine learning
Networks
Sampling
title Paused Agent Replay Refresh
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2025-01-08T09%3A20%3A52IST&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=Paused%20Agent%20Replay%20Refresh&rft.jtitle=arXiv.org&rft.au=Parr,%20Benjamin&rft.date=2022-09-26&rft.eissn=2331-8422&rft_id=info:doi/&rft_dat=%3Cproquest%3E2718739415%3C/proquest%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=2718739415&rft_id=info:pmid/&rfr_iscdi=true