{"id":"https://openalex.org/W3114874950","doi":"https://doi.org/10.1109/hpec43674.2020.9286150","title":"How to Efficiently Train Your AI Agent? Characterizing and Evaluating Deep Reinforcement Learning on Heterogeneous Platforms","display_name":"How to Efficiently Train Your AI Agent? Characterizing and Evaluating Deep Reinforcement Learning on Heterogeneous Platforms","publication_year":2020,"publication_date":"2020-09-22","ids":{"openalex":"https://openalex.org/W3114874950","doi":"https://doi.org/10.1109/hpec43674.2020.9286150","mag":"3114874950"},"language":"en","primary_location":{"id":"doi:10.1109/hpec43674.2020.9286150","is_oa":false,"landing_page_url":"https://doi.org/10.1109/hpec43674.2020.9286150","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2020 IEEE High Performance Extreme Computing Conference (HPEC)","raw_type":"proceedings-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":false,"oa_status":"closed","oa_url":null,"any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5050528589","display_name":"Yuan Meng","orcid":"https://orcid.org/0000-0001-6468-8623"},"institutions":[{"id":"https://openalex.org/I2800817003","display_name":"Southern California University for Professional Studies","ror":"https://ror.org/058zz0t50","country_code":"US","type":"education","lineage":["https://openalex.org/I2800817003"]},{"id":"https://openalex.org/I1174212","display_name":"University of Southern California","ror":"https://ror.org/03taz7m60","country_code":"US","type":"education","lineage":["https://openalex.org/I1174212"]}],"countries":["US"],"is_corresponding":true,"raw_author_name":"Yuan Meng","raw_affiliation_strings":["University of Southern California"],"affiliations":[{"raw_affiliation_string":"University of Southern California","institution_ids":["https://openalex.org/I2800817003","https://openalex.org/I1174212"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100397725","display_name":"Yang Yang","orcid":"https://orcid.org/0000-0003-0608-9408"},"institutions":[{"id":"https://openalex.org/I2800817003","display_name":"Southern California University for Professional Studies","ror":"https://ror.org/058zz0t50","country_code":"US","type":"education","lineage":["https://openalex.org/I2800817003"]},{"id":"https://openalex.org/I1174212","display_name":"University of Southern California","ror":"https://ror.org/03taz7m60","country_code":"US","type":"education","lineage":["https://openalex.org/I1174212"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Yang Yang","raw_affiliation_strings":["University of Southern California"],"affiliations":[{"raw_affiliation_string":"University of Southern California","institution_ids":["https://openalex.org/I2800817003","https://openalex.org/I1174212"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5015036207","display_name":"Sanmukh R. Kuppannagari","orcid":"https://orcid.org/0000-0002-2062-1483"},"institutions":[{"id":"https://openalex.org/I2800817003","display_name":"Southern California University for Professional Studies","ror":"https://ror.org/058zz0t50","country_code":"US","type":"education","lineage":["https://openalex.org/I2800817003"]},{"id":"https://openalex.org/I1174212","display_name":"University of Southern California","ror":"https://ror.org/03taz7m60","country_code":"US","type":"education","lineage":["https://openalex.org/I1174212"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Sanmukh Kuppannagari","raw_affiliation_strings":["University of Southern California"],"affiliations":[{"raw_affiliation_string":"University of Southern California","institution_ids":["https://openalex.org/I2800817003","https://openalex.org/I1174212"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5042560222","display_name":"Rajgopal Kannan","orcid":"https://orcid.org/0000-0001-8736-3012"},"institutions":[{"id":"https://openalex.org/I166416128","display_name":"DEVCOM Army Research Laboratory","ror":"https://ror.org/011hc8f90","country_code":"US","type":"government","lineage":["https://openalex.org/I1304082316","https://openalex.org/I1330347796","https://openalex.org/I166416128","https://openalex.org/I2802705668","https://openalex.org/I4210154437"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Rajgopal Kannan","raw_affiliation_strings":["US Army Research Lab"],"affiliations":[{"raw_affiliation_string":"US Army Research Lab","institution_ids":["https://openalex.org/I166416128"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5033166029","display_name":"Viktor K. Prasanna","orcid":"https://orcid.org/0000-0002-1609-8589"},"institutions":[{"id":"https://openalex.org/I2800817003","display_name":"Southern California University for Professional Studies","ror":"https://ror.org/058zz0t50","country_code":"US","type":"education","lineage":["https://openalex.org/I2800817003"]},{"id":"https://openalex.org/I1174212","display_name":"University of Southern California","ror":"https://ror.org/03taz7m60","country_code":"US","type":"education","lineage":["https://openalex.org/I1174212"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Viktor Prasanna","raw_affiliation_strings":["University of Southern California"],"affiliations":[{"raw_affiliation_string":"University of Southern California","institution_ids":["https://openalex.org/I2800817003","https://openalex.org/I1174212"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":5,"corresponding_author_ids":["https://openalex.org/A5050528589"],"corresponding_institution_ids":["https://openalex.org/I1174212","https://openalex.org/I2800817003"],"apc_list":null,"apc_paid":null,"fwci":0.5302,"has_fulltext":false,"cited_by_count":5,"citation_normalized_percentile":{"value":0.7391775,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":89,"max":96},"biblio":{"volume":null,"issue":null,"first_page":"1","last_page":"7"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10462","display_name":"Reinforcement Learning in Robotics","score":0.9995999932289124,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},"topics":[{"id":"https://openalex.org/T10462","display_name":"Reinforcement Learning in Robotics","score":0.9995999932289124,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.9944000244140625,"subfield":{"id":"https://openalex.org/subfields/1702","display_name":"Artificial Intelligence"},"field":{"id":"https://openalex.org/fields/17","display_name":"Computer Science"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T11099","display_name":"Autonomous Vehicle Technology and Safety","score":0.9710000157356262,"subfield":{"id":"https://openalex.org/subfields/2203","display_name":"Automotive Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.8289669752120972},{"id":"https://openalex.org/keywords/reinforcement-learning","display_name":"Reinforcement learning","score":0.7922048568725586},{"id":"https://openalex.org/keywords/deep-learning","display_name":"Deep learning","score":0.7667300701141357},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.6100322008132935},{"id":"https://openalex.org/keywords/deep-neural-networks","display_name":"Deep neural networks","score":0.5496646761894226},{"id":"https://openalex.org/keywords/field-programmable-gate-array","display_name":"Field-programmable gate array","score":0.5492909550666809},{"id":"https://openalex.org/keywords/memory-hierarchy","display_name":"Memory hierarchy","score":0.4830766022205353},{"id":"https://openalex.org/keywords/computation","display_name":"Computation","score":0.4750397503376007},{"id":"https://openalex.org/keywords/abstraction","display_name":"Abstraction","score":0.4742695093154907},{"id":"https://openalex.org/keywords/robotics","display_name":"Robotics","score":0.47145992517471313},{"id":"https://openalex.org/keywords/latency","display_name":"Latency (audio)","score":0.4514363408088684},{"id":"https://openalex.org/keywords/artificial-neural-network","display_name":"Artificial neural network","score":0.4385712742805481},{"id":"https://openalex.org/keywords/architecture","display_name":"Architecture","score":0.4382176399230957},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.43474018573760986},{"id":"https://openalex.org/keywords/distributed-computing","display_name":"Distributed computing","score":0.38172677159309387},{"id":"https://openalex.org/keywords/computer-architecture","display_name":"Computer architecture","score":0.3638356328010559},{"id":"https://openalex.org/keywords/computer-engineering","display_name":"Computer engineering","score":0.34901514649391174},{"id":"https://openalex.org/keywords/embedded-system","display_name":"Embedded system","score":0.2916181683540344},{"id":"https://openalex.org/keywords/parallel-computing","display_name":"Parallel computing","score":0.27201908826828003},{"id":"https://openalex.org/keywords/robot","display_name":"Robot","score":0.2148299217224121},{"id":"https://openalex.org/keywords/algorithm","display_name":"Algorithm","score":0.15468239784240723}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8289669752120972},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.7922048568725586},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.7667300701141357},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.6100322008132935},{"id":"https://openalex.org/C2984842247","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep neural networks","level":3,"score":0.5496646761894226},{"id":"https://openalex.org/C42935608","wikidata":"https://www.wikidata.org/wiki/Q190411","display_name":"Field-programmable gate array","level":2,"score":0.5492909550666809},{"id":"https://openalex.org/C2778100165","wikidata":"https://www.wikidata.org/wiki/Q1589327","display_name":"Memory hierarchy","level":3,"score":0.4830766022205353},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.4750397503376007},{"id":"https://openalex.org/C124304363","wikidata":"https://www.wikidata.org/wiki/Q673661","display_name":"Abstraction","level":2,"score":0.4742695093154907},{"id":"https://openalex.org/C34413123","wikidata":"https://www.wikidata.org/wiki/Q170978","display_name":"Robotics","level":3,"score":0.47145992517471313},{"id":"https://openalex.org/C82876162","wikidata":"https://www.wikidata.org/wiki/Q17096504","display_name":"Latency (audio)","level":2,"score":0.4514363408088684},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.4385712742805481},{"id":"https://openalex.org/C123657996","wikidata":"https://www.wikidata.org/wiki/Q12271","display_name":"Architecture","level":2,"score":0.4382176399230957},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.43474018573760986},{"id":"https://openalex.org/C120314980","wikidata":"https://www.wikidata.org/wiki/Q180634","display_name":"Distributed computing","level":1,"score":0.38172677159309387},{"id":"https://openalex.org/C118524514","wikidata":"https://www.wikidata.org/wiki/Q173212","display_name":"Computer architecture","level":1,"score":0.3638356328010559},{"id":"https://openalex.org/C113775141","wikidata":"https://www.wikidata.org/wiki/Q428691","display_name":"Computer engineering","level":1,"score":0.34901514649391174},{"id":"https://openalex.org/C149635348","wikidata":"https://www.wikidata.org/wiki/Q193040","display_name":"Embedded system","level":1,"score":0.2916181683540344},{"id":"https://openalex.org/C173608175","wikidata":"https://www.wikidata.org/wiki/Q232661","display_name":"Parallel computing","level":1,"score":0.27201908826828003},{"id":"https://openalex.org/C90509273","wikidata":"https://www.wikidata.org/wiki/Q11012","display_name":"Robot","level":2,"score":0.2148299217224121},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.15468239784240723},{"id":"https://openalex.org/C115537543","wikidata":"https://www.wikidata.org/wiki/Q165596","display_name":"Cache","level":2,"score":0.0},{"id":"https://openalex.org/C76155785","wikidata":"https://www.wikidata.org/wiki/Q418","display_name":"Telecommunications","level":1,"score":0.0},{"id":"https://openalex.org/C153349607","wikidata":"https://www.wikidata.org/wiki/Q36649","display_name":"Visual arts","level":1,"score":0.0},{"id":"https://openalex.org/C38652104","wikidata":"https://www.wikidata.org/wiki/Q3510521","display_name":"Computer security","level":1,"score":0.0},{"id":"https://openalex.org/C142362112","wikidata":"https://www.wikidata.org/wiki/Q735","display_name":"Art","level":0,"score":0.0},{"id":"https://openalex.org/C111472728","wikidata":"https://www.wikidata.org/wiki/Q9471","display_name":"Epistemology","level":1,"score":0.0},{"id":"https://openalex.org/C138885662","wikidata":"https://www.wikidata.org/wiki/Q5891","display_name":"Philosophy","level":0,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/hpec43674.2020.9286150","is_oa":false,"landing_page_url":"https://doi.org/10.1109/hpec43674.2020.9286150","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2020 IEEE High Performance Extreme Computing Conference (HPEC)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":33,"referenced_works":["https://openalex.org/W1510543252","https://openalex.org/W2065087844","https://openalex.org/W2145339207","https://openalex.org/W2145800860","https://openalex.org/W2158782408","https://openalex.org/W2173248099","https://openalex.org/W2201581102","https://openalex.org/W2342662072","https://openalex.org/W2523246573","https://openalex.org/W2557443516","https://openalex.org/W2576404523","https://openalex.org/W2736601468","https://openalex.org/W2786036274","https://openalex.org/W2796649226","https://openalex.org/W2811374584","https://openalex.org/W2931767035","https://openalex.org/W2950471160","https://openalex.org/W2953384591","https://openalex.org/W2953708620","https://openalex.org/W2963094322","https://openalex.org/W2963477884","https://openalex.org/W2963641140","https://openalex.org/W2963864421","https://openalex.org/W2964043796","https://openalex.org/W2964044380","https://openalex.org/W3035681682","https://openalex.org/W4246367117","https://openalex.org/W4297797010","https://openalex.org/W4301501993","https://openalex.org/W6692846177","https://openalex.org/W6696324988","https://openalex.org/W6704571135","https://openalex.org/W6713134421"],"related_works":["https://openalex.org/W2045155990","https://openalex.org/W4313163053","https://openalex.org/W4300973204","https://openalex.org/W3045811229","https://openalex.org/W1508899372","https://openalex.org/W1483408780","https://openalex.org/W4284884309","https://openalex.org/W2908749798","https://openalex.org/W2105980483","https://openalex.org/W4285818394"],"abstract_inverted_index":{"Deep":[0,20,24,47,76,112,143,156,185,217,222,236],"Reinforcement":[1],"Learning":[2],"(Deep":[3],"RL)":[4],"is":[5,54],"a":[6,23,39,46,51,175,181,206,211,227,263],"key":[7],"technology":[8],"in":[9,111,190],"several":[10,57],"domains":[11],"such":[12],"as":[13],"self-driving":[14],"cars,":[15],"robotics,":[16],"surveillance,":[17],"etc.":[18],"In":[19,166],"RL,":[21],"using":[22],"Neural":[25],"Network":[26],"model,":[27],"an":[28],"agent":[29],"learns":[30],"how":[31],"to":[32,37,75,87,99,126,152,269],"interact":[33],"with":[34],"the":[35,61,64,79,82,93,96,104,129,153,171,231],"environment":[36],"achieve":[38],"certain":[40],"goal.":[41],"The":[42],"efficiency":[43],"of":[44,63,81,95,155,173,184,216,230,233],"running":[45],"RL":[48,113,121,144,157,186,218,237],"algorithm":[49],"on":[50,239,276,287],"hardware":[52,65,83,97],"architecture":[53,66,98],"dependent":[55],"upon":[56],"factors":[58],"including":[59,188],"(1)":[60],"suitability":[62],"for":[67,119,141,210,253],"kernels":[68],"and":[69,91,159,241,251,258,272,279,282],"computation":[70,109],"patterns":[71],"which":[72],"are":[73,164],"fundamental":[74],"RL;":[77],"(2)":[78],"capability":[80],"architecture's":[84],"memory":[85],"hierarchy":[86],"minimize":[88],"data-communication":[89],"latency;":[90],"(3)":[92],"ability":[94],"hide":[100],"overheads":[101],"introduced":[102],"by":[103,204],"deeply":[105],"nested":[106],"highly":[107,138],"irregular":[108],"characteristics":[110],"algorithms.":[114,145],"GPUs":[115],"have":[116,136],"been":[117],"popular":[118],"accelerating":[120],"algorithms,":[122],"however,":[123],"they":[124,147],"fail":[125],"optimally":[127],"satisfy":[128],"above-mentioned":[130],"requirements.":[131],"A":[132],"few":[133],"recent":[134],"works":[135],"developed":[137],"customized":[139],"accelerators":[140],"specific":[142],"However,":[146],"cannot":[148],"be":[149],"generalized":[150],"easily":[151],"plethora":[154],"algorithms":[158,187,219,238,248],"DNN":[160,194],"model":[161,195],"choices":[162],"that":[163,178,262],"available.":[165],"this":[167,202],"paper,":[168],"we":[169,225,260],"explore":[170],"possibility":[172],"developing":[174],"unified":[176],"framework":[177],"can":[179],"accelerate":[180],"wide":[182],"range":[183],"variations":[189],"training":[191,277],"methods":[192],"or":[193],"structures.":[196],"We":[197,244],"take":[198],"one":[199],"step":[200],"towards":[201],"goal":[203],"defining":[205],"domain-specific":[207],"high-level":[208],"abstraction":[209],"widely":[212],"used":[213],"broad":[214],"class":[215],"-":[220,249,256],"on-policy":[221,235],"RL.":[223],"Furthermore,":[224],"provide":[226],"systematic":[228],"analysis":[229],"performance":[232],"state-of-the-art":[234],"CPU-GPU":[240],"CPU-FPGA":[242],"platforms.":[243],"target":[245],"two":[246],"representative":[247],"PPO":[250],"A2C,":[252],"application":[254],"areas":[255],"robotics":[257],"games.":[259],"show":[261],"FPGA-based":[264],"custom":[265],"accelerator":[266],"achieves":[267],"up":[268],"24\u00d7":[270],"(PPO)":[271,281],"8\u00d7":[273],"(A2C)":[274,285],"speedups":[275],"tasks,":[278],"17\u00d7":[280],"2.1":[283],"\u00d7":[284],"improvements":[286],"overall":[288],"throughput,":[289],"respectively.":[290]},"counts_by_year":[{"year":2025,"cited_by_count":1},{"year":2023,"cited_by_count":2},{"year":2022,"cited_by_count":1},{"year":2021,"cited_by_count":1}],"updated_date":"2025-11-06T03:46:38.306776","created_date":"2025-10-10T00:00:00"}