ARLBench: Flexible and Efficient Benchmarking for Hyperparameter Optimization in Reinforcement Learning

Hyperparameters are a critical factor in reliably training well-performing reinforcement learning (RL) agents. Unfortunately, developing and evaluating automated approaches for tuning such hyperparameters is both costly and time-consuming. As a result, such approaches are often only evaluated on a s...

Ausführliche Beschreibung

Gespeichert in:
Bibliographische Detailangaben
Veröffentlicht in:arXiv.org 2024-09
Hauptverfasser: Becktepe, Jannis, Dierkes, Julian, Benjamins, Carolin, Mohan, Aditya, Salinas, David, Rajan, Raghu, Hutter, Frank, Hoos, Holger, Lindauer, Marius, Eimer, Theresa
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 Becktepe, Jannis
Dierkes, Julian
Benjamins, Carolin
Mohan, Aditya
Salinas, David
Rajan, Raghu
Hutter, Frank
Hoos, Holger
Lindauer, Marius
Eimer, Theresa
description Hyperparameters are a critical factor in reliably training well-performing reinforcement learning (RL) agents. Unfortunately, developing and evaluating automated approaches for tuning such hyperparameters is both costly and time-consuming. As a result, such approaches are often only evaluated on a single domain or algorithm, making comparisons difficult and limiting insights into their generalizability. We propose ARLBench, a benchmark for hyperparameter optimization (HPO) in RL that allows comparisons of diverse HPO approaches while being highly efficient in evaluation. To enable research into HPO in RL, even in settings with low compute resources, we select a representative subset of HPO tasks spanning a variety of algorithm and environment combinations. This selection allows for generating a performance profile of an automated RL (AutoRL) method using only a fraction of the compute previously necessary, enabling a broader range of researchers to work on HPO in RL. With the extensive and large-scale dataset on hyperparameter landscapes that our selection is based on, ARLBench is an efficient, flexible, and future-oriented foundation for research on AutoRL. Both the benchmark and the dataset are available at https://github.com/automl/arlbench.
format Article
fullrecord <record><control><sourceid>proquest</sourceid><recordid>TN_cdi_proquest_journals_3111340698</recordid><sourceformat>XML</sourceformat><sourcesystem>PC</sourcesystem><sourcerecordid>3111340698</sourcerecordid><originalsourceid>FETCH-proquest_journals_31113406983</originalsourceid><addsrcrecordid>eNqNi8EKgkAUAJcgKMp_eNA50F0z61aRdAiC6B6bPe2VPm13herrs-gDOs1hZjqiL5UKxnEoZU941l5935fRVE4mqi_yxX67RE4vc0gKfNCpQNB8hnWWUUrIDr621OZGnENWGdg8azS1NrpEhwZ2taOSXtpRxUAMeyRusxTLz71Fbbg9h6Kb6cKi9-NAjJL1YbUZ16a6N2jd8Vo1hlt1VEEQqNCPZrH6r3oD5hhHkA</addsrcrecordid><sourcetype>Aggregation Database</sourcetype><iscdi>true</iscdi><recordtype>article</recordtype><pqid>3111340698</pqid></control><display><type>article</type><title>ARLBench: Flexible and Efficient Benchmarking for Hyperparameter Optimization in Reinforcement Learning</title><source>Free E- Journals</source><creator>Becktepe, Jannis ; Dierkes, Julian ; Benjamins, Carolin ; Mohan, Aditya ; Salinas, David ; Rajan, Raghu ; Hutter, Frank ; Hoos, Holger ; Lindauer, Marius ; Eimer, Theresa</creator><creatorcontrib>Becktepe, Jannis ; Dierkes, Julian ; Benjamins, Carolin ; Mohan, Aditya ; Salinas, David ; Rajan, Raghu ; Hutter, Frank ; Hoos, Holger ; Lindauer, Marius ; Eimer, Theresa</creatorcontrib><description>Hyperparameters are a critical factor in reliably training well-performing reinforcement learning (RL) agents. Unfortunately, developing and evaluating automated approaches for tuning such hyperparameters is both costly and time-consuming. As a result, such approaches are often only evaluated on a single domain or algorithm, making comparisons difficult and limiting insights into their generalizability. We propose ARLBench, a benchmark for hyperparameter optimization (HPO) in RL that allows comparisons of diverse HPO approaches while being highly efficient in evaluation. To enable research into HPO in RL, even in settings with low compute resources, we select a representative subset of HPO tasks spanning a variety of algorithm and environment combinations. This selection allows for generating a performance profile of an automated RL (AutoRL) method using only a fraction of the compute previously necessary, enabling a broader range of researchers to work on HPO in RL. With the extensive and large-scale dataset on hyperparameter landscapes that our selection is based on, ARLBench is an efficient, flexible, and future-oriented foundation for research on AutoRL. Both the benchmark and the dataset are available at https://github.com/automl/arlbench.</description><identifier>EISSN: 2331-8422</identifier><language>eng</language><publisher>Ithaca: Cornell University Library, arXiv.org</publisher><subject>Algorithms ; Automation ; Benchmarks ; Datasets ; Optimization</subject><ispartof>arXiv.org, 2024-09</ispartof><rights>2024. This work is published under http://creativecommons.org/licenses/by/4.0/ (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>781,785</link.rule.ids></links><search><creatorcontrib>Becktepe, Jannis</creatorcontrib><creatorcontrib>Dierkes, Julian</creatorcontrib><creatorcontrib>Benjamins, Carolin</creatorcontrib><creatorcontrib>Mohan, Aditya</creatorcontrib><creatorcontrib>Salinas, David</creatorcontrib><creatorcontrib>Rajan, Raghu</creatorcontrib><creatorcontrib>Hutter, Frank</creatorcontrib><creatorcontrib>Hoos, Holger</creatorcontrib><creatorcontrib>Lindauer, Marius</creatorcontrib><creatorcontrib>Eimer, Theresa</creatorcontrib><title>ARLBench: Flexible and Efficient Benchmarking for Hyperparameter Optimization in Reinforcement Learning</title><title>arXiv.org</title><description>Hyperparameters are a critical factor in reliably training well-performing reinforcement learning (RL) agents. Unfortunately, developing and evaluating automated approaches for tuning such hyperparameters is both costly and time-consuming. As a result, such approaches are often only evaluated on a single domain or algorithm, making comparisons difficult and limiting insights into their generalizability. We propose ARLBench, a benchmark for hyperparameter optimization (HPO) in RL that allows comparisons of diverse HPO approaches while being highly efficient in evaluation. To enable research into HPO in RL, even in settings with low compute resources, we select a representative subset of HPO tasks spanning a variety of algorithm and environment combinations. This selection allows for generating a performance profile of an automated RL (AutoRL) method using only a fraction of the compute previously necessary, enabling a broader range of researchers to work on HPO in RL. With the extensive and large-scale dataset on hyperparameter landscapes that our selection is based on, ARLBench is an efficient, flexible, and future-oriented foundation for research on AutoRL. Both the benchmark and the dataset are available at https://github.com/automl/arlbench.</description><subject>Algorithms</subject><subject>Automation</subject><subject>Benchmarks</subject><subject>Datasets</subject><subject>Optimization</subject><issn>2331-8422</issn><fulltext>true</fulltext><rsrctype>article</rsrctype><creationdate>2024</creationdate><recordtype>article</recordtype><sourceid>ABUWG</sourceid><sourceid>AFKRA</sourceid><sourceid>AZQEC</sourceid><sourceid>BENPR</sourceid><sourceid>CCPQU</sourceid><sourceid>DWQXO</sourceid><recordid>eNqNi8EKgkAUAJcgKMp_eNA50F0z61aRdAiC6B6bPe2VPm13herrs-gDOs1hZjqiL5UKxnEoZU941l5935fRVE4mqi_yxX67RE4vc0gKfNCpQNB8hnWWUUrIDr621OZGnENWGdg8azS1NrpEhwZ2taOSXtpRxUAMeyRusxTLz71Fbbg9h6Kb6cKi9-NAjJL1YbUZ16a6N2jd8Vo1hlt1VEEQqNCPZrH6r3oD5hhHkA</recordid><startdate>20240927</startdate><enddate>20240927</enddate><creator>Becktepe, Jannis</creator><creator>Dierkes, Julian</creator><creator>Benjamins, Carolin</creator><creator>Mohan, Aditya</creator><creator>Salinas, David</creator><creator>Rajan, Raghu</creator><creator>Hutter, Frank</creator><creator>Hoos, Holger</creator><creator>Lindauer, Marius</creator><creator>Eimer, Theresa</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>20240927</creationdate><title>ARLBench: Flexible and Efficient Benchmarking for Hyperparameter Optimization in Reinforcement Learning</title><author>Becktepe, Jannis ; Dierkes, Julian ; Benjamins, Carolin ; Mohan, Aditya ; Salinas, David ; Rajan, Raghu ; Hutter, Frank ; Hoos, Holger ; Lindauer, Marius ; Eimer, Theresa</author></sort><facets><frbrtype>5</frbrtype><frbrgroupid>cdi_FETCH-proquest_journals_31113406983</frbrgroupid><rsrctype>articles</rsrctype><prefilter>articles</prefilter><language>eng</language><creationdate>2024</creationdate><topic>Algorithms</topic><topic>Automation</topic><topic>Benchmarks</topic><topic>Datasets</topic><topic>Optimization</topic><toplevel>online_resources</toplevel><creatorcontrib>Becktepe, Jannis</creatorcontrib><creatorcontrib>Dierkes, Julian</creatorcontrib><creatorcontrib>Benjamins, Carolin</creatorcontrib><creatorcontrib>Mohan, Aditya</creatorcontrib><creatorcontrib>Salinas, David</creatorcontrib><creatorcontrib>Rajan, Raghu</creatorcontrib><creatorcontrib>Hutter, Frank</creatorcontrib><creatorcontrib>Hoos, Holger</creatorcontrib><creatorcontrib>Lindauer, Marius</creatorcontrib><creatorcontrib>Eimer, Theresa</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>Becktepe, Jannis</au><au>Dierkes, Julian</au><au>Benjamins, Carolin</au><au>Mohan, Aditya</au><au>Salinas, David</au><au>Rajan, Raghu</au><au>Hutter, Frank</au><au>Hoos, Holger</au><au>Lindauer, Marius</au><au>Eimer, Theresa</au><format>book</format><genre>document</genre><ristype>GEN</ristype><atitle>ARLBench: Flexible and Efficient Benchmarking for Hyperparameter Optimization in Reinforcement Learning</atitle><jtitle>arXiv.org</jtitle><date>2024-09-27</date><risdate>2024</risdate><eissn>2331-8422</eissn><abstract>Hyperparameters are a critical factor in reliably training well-performing reinforcement learning (RL) agents. Unfortunately, developing and evaluating automated approaches for tuning such hyperparameters is both costly and time-consuming. As a result, such approaches are often only evaluated on a single domain or algorithm, making comparisons difficult and limiting insights into their generalizability. We propose ARLBench, a benchmark for hyperparameter optimization (HPO) in RL that allows comparisons of diverse HPO approaches while being highly efficient in evaluation. To enable research into HPO in RL, even in settings with low compute resources, we select a representative subset of HPO tasks spanning a variety of algorithm and environment combinations. This selection allows for generating a performance profile of an automated RL (AutoRL) method using only a fraction of the compute previously necessary, enabling a broader range of researchers to work on HPO in RL. With the extensive and large-scale dataset on hyperparameter landscapes that our selection is based on, ARLBench is an efficient, flexible, and future-oriented foundation for research on AutoRL. Both the benchmark and the dataset are available at https://github.com/automl/arlbench.</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, 2024-09
issn 2331-8422
language eng
recordid cdi_proquest_journals_3111340698
source Free E- Journals
subjects Algorithms
Automation
Benchmarks
Datasets
Optimization
title ARLBench: Flexible and Efficient Benchmarking for Hyperparameter Optimization in Reinforcement Learning
url https://sfx.bib-bvb.de/sfx_tum?ctx_ver=Z39.88-2004&ctx_enc=info:ofi/enc:UTF-8&ctx_tim=2024-12-18T08%3A03%3A56IST&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=ARLBench:%20Flexible%20and%20Efficient%20Benchmarking%20for%20Hyperparameter%20Optimization%20in%20Reinforcement%20Learning&rft.jtitle=arXiv.org&rft.au=Becktepe,%20Jannis&rft.date=2024-09-27&rft.eissn=2331-8422&rft_id=info:doi/&rft_dat=%3Cproquest%3E3111340698%3C/proquest%3E%3Curl%3E%3C/url%3E&disable_directlink=true&sfx.directlink=off&sfx.report_link=0&rft_id=info:oai/&rft_pqid=3111340698&rft_id=info:pmid/&rfr_iscdi=true