{"id":"https://openalex.org/W2927749396","doi":null,"title":"Samples are not all useful: Denoising policy gradient updates using variance","display_name":"Samples are not all useful: Denoising policy gradient updates using variance","publication_year":2019,"publication_date":"2019-04-10","ids":{"openalex":"https://openalex.org/W2927749396","mag":"2927749396"},"language":"en","primary_location":{"id":"pmh:oai:HAL:hal-02091547v2","doi":null,"is_oa":true,"landing_page_url":"https://hal.inria.fr/hal-02091547","pdf_url":null,"source":{"id":"https://openalex.org/S4306402512","display_name":"HAL","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I4405252976","host_organization_name":"Centre pour la Communication Scientifique Directe","host_organization_lineage":["https://openalex.org/I4405252976"],"host_organization_lineage_names":[],"type":"repository"},"license":"other-oa","license_id":"https://openalex.org/licenses/other-oa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"https://hal.inria.fr/hal-02091547","raw_type":"info:eu-repo/semantics/preprint"},"volume":null,"type":"preprint","is_traditional":true,"indexed_in":[],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://hal.inria.fr/hal-02091547","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5045192384","display_name":"Yannis Flet-Berliac","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I4387153239","display_name":"Centre de Recherche en Informatique, Signal et Automatique de Lille","ror":"https://ror.org/05vrs3189","country_code":"FR","type":"facility","lineage":["https://openalex.org/I1294671590","https://openalex.org/I1326498283","https://openalex.org/I2279609970","https://openalex.org/I4387153239","https://openalex.org/I7454413"]}],"countries":["FR"],"is_corresponding":false,"raw_author_name":"Flet-Berliac, Yannis","raw_affiliation_strings":["Centre de Recherche en Informatique, Signal et Automatique de Lille - UMR 9189","Scool","Sequential Learning"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Centre de Recherche en Informatique, Signal et Automatique de Lille - UMR 9189","institution_ids":["https://openalex.org/I4387153239"]},{"raw_affiliation_string":"Scool","institution_ids":[]},{"raw_affiliation_string":"Sequential Learning","institution_ids":[]}]},{"author_position":"last","author":{"id":null,"display_name":"Preux, Philippe","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Preux, Philippe","raw_affiliation_strings":["Scool","Sequential Learning"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Scool","institution_ids":[]},{"raw_affiliation_string":"Sequential Learning","institution_ids":[]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"cited_by_count":2,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":null,"last_page":null},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10462","display_name":"Reinforcement Learning in Robotics","score":1.0,"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":1.0,"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/T12794","display_name":"Adaptive Dynamic Programming Control","score":0.0,"subfield":{"id":"https://openalex.org/subfields/2207","display_name":"Control and Systems Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T11195","display_name":"Simulation Techniques and Applications","score":0.0,"subfield":{"id":"https://openalex.org/subfields/1803","display_name":"Management Science and Operations Research"},"field":{"id":"https://openalex.org/fields/18","display_name":"Decision Sciences"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/policy-gradient","display_name":"policy gradient","score":0.9240000247955322},{"id":"https://openalex.org/keywords/sample-selection","display_name":"sample selection","score":0.5450000166893005},{"id":"https://openalex.org/keywords/explained-variance","display_name":"explained variance","score":0.8100000023841858},{"id":"https://openalex.org/keywords/sampling-efficiency","display_name":"sampling efficiency","score":0.48899999260902405},{"id":"https://openalex.org/keywords/exploration-exploitation-trade-off","display_name":"exploration-exploitation balance","score":0.9100000262260437},{"id":"https://openalex.org/keywords/continuous-control","display_name":"continuous control","score":0.6230000257492065},{"id":"https://openalex.org/keywords/reinforcement-learning","display_name":"reinforcement learning","score":0.6330000162124634},{"id":"https://openalex.org/keywords/model-based-sampling","display_name":"model-based sampling","score":0.32899999618530273}],"concepts":[{"id":"https://openalex.org/C196083921","wikidata":"https://www.wikidata.org/wiki/Q7915758","display_name":"Variance (accounting)","level":2,"score":0.6253453493118286},{"id":"https://openalex.org/C163294075","wikidata":"https://www.wikidata.org/wiki/Q581861","display_name":"Noise reduction","level":2,"score":0.5442635416984558},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5091633200645447},{"id":"https://openalex.org/C62644790","wikidata":"https://www.wikidata.org/wiki/Q3454689","display_name":"Variance reduction","level":3,"score":0.46520763635635376},{"id":"https://openalex.org/C149782125","wikidata":"https://www.wikidata.org/wiki/Q160039","display_name":"Econometrics","level":1,"score":0.3906317949295044},{"id":"https://openalex.org/C105795698","wikidata":"https://www.wikidata.org/wiki/Q12483","display_name":"Statistics","level":1,"score":0.3754056692123413},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3180930018424988},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.27439865469932556},{"id":"https://openalex.org/C144133560","wikidata":"https://www.wikidata.org/wiki/Q4830453","display_name":"Business","level":0,"score":0.1883309781551361},{"id":"https://openalex.org/C121955636","wikidata":"https://www.wikidata.org/wiki/Q4116214","display_name":"Accounting","level":1,"score":0.057444602251052856}],"mesh":[],"locations_count":1,"locations":[{"id":"pmh:oai:HAL:hal-02091547v2","doi":null,"is_oa":true,"landing_page_url":"https://hal.inria.fr/hal-02091547","pdf_url":null,"source":{"id":"https://openalex.org/S4306402512","display_name":"HAL","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I4405252976","host_organization_name":"Centre pour la Communication Scientifique Directe","host_organization_lineage":["https://openalex.org/I4405252976"],"host_organization_lineage_names":[],"type":"repository"},"license":"other-oa","license_id":"https://openalex.org/licenses/other-oa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"https://hal.inria.fr/hal-02091547","raw_type":"info:eu-repo/semantics/preprint"}],"best_oa_location":{"id":"pmh:oai:HAL:hal-02091547v2","doi":null,"is_oa":true,"landing_page_url":"https://hal.inria.fr/hal-02091547","pdf_url":null,"source":{"id":"https://openalex.org/S4306402512","display_name":"HAL","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I4405252976","host_organization_name":"Centre pour la Communication Scientifique Directe","host_organization_lineage":["https://openalex.org/I4405252976"],"host_organization_lineage_names":[],"type":"repository"},"license":"other-oa","license_id":"https://openalex.org/licenses/other-oa","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":"https://hal.inria.fr/hal-02091547","raw_type":"info:eu-repo/semantics/preprint"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":44,"referenced_works":["https://openalex.org/W107583932","https://openalex.org/W1515851193","https://openalex.org/W1583833196","https://openalex.org/W1585861384","https://openalex.org/W1757796397","https://openalex.org/W1863227302","https://openalex.org/W1904826605","https://openalex.org/W1906772730","https://openalex.org/W1907796993","https://openalex.org/W2057327562","https://openalex.org/W2095705004","https://openalex.org/W2109910161","https://openalex.org/W2119567691","https://openalex.org/W2121103318","https://openalex.org/W2121863487","https://openalex.org/W2125612430","https://openalex.org/W2130801532","https://openalex.org/W2150468603","https://openalex.org/W2155027007","https://openalex.org/W2158782408","https://openalex.org/W2165150801","https://openalex.org/W2173248099","https://openalex.org/W2201581102","https://openalex.org/W2257979135","https://openalex.org/W2260756217","https://openalex.org/W2556958149","https://openalex.org/W2593044849","https://openalex.org/W2594829461","https://openalex.org/W2726187156","https://openalex.org/W2736601468","https://openalex.org/W2749928749","https://openalex.org/W2761873684","https://openalex.org/W2781726626","https://openalex.org/W2786036274","https://openalex.org/W2804901184","https://openalex.org/W2809668646","https://openalex.org/W2890208753","https://openalex.org/W2898436992","https://openalex.org/W2899205164","https://openalex.org/W2903181768","https://openalex.org/W2949608212","https://openalex.org/W2963184621","https://openalex.org/W3130385304","https://openalex.org/W4285719527"],"related_works":["https://openalex.org/W3026994642","https://openalex.org/W2133224499","https://openalex.org/W2338705551","https://openalex.org/W3006333061","https://openalex.org/W2766436268","https://openalex.org/W7221765543","https://openalex.org/W3149322806","https://openalex.org/W370713089","https://openalex.org/W1596364083","https://openalex.org/W2976021111"],"abstract_inverted_index":{"Policy":[0],"gradient":[1,61,115,167],"algorithms":[2,30],"in":[3,34,59],"reinforcement":[4],"learning":[5],"rely":[6],"on":[7,18,104],"efficiently":[8],"sampling":[9,13,19,39,110,133],"an":[10],"environment.":[11],"Most":[12],"procedures":[14],"are":[15],"based":[16],"solely":[17],"the":[20,38,60,75,84,98,105,113,126,132,161,165,179,187,204],"agent's":[21],"policy.":[22],"However,":[23],"other":[24],"measures":[25],"made":[26],"available":[27],"through":[28],"these":[29],"could":[31],"be":[32,94,196],"used":[33,58,95,129],"order":[35],"to":[36,41,96,130,202],"improve":[37,131],"prior":[40],"each":[42,100],"policy":[43,114,166],"update.":[44,168],"Following":[45],"this":[46,118,193],"line":[47],"of":[48,77,83,142,181],"thoughts,":[49],"we":[50,157],"propose":[51],"a":[52,55,66,81,87,140,175,199],"method":[53],"where":[54],"transition":[56,101],"is":[57,74],"update":[62],"if":[63,156],"it":[64],"meets":[65],"particular":[67],"criterion,":[68],"and":[69,89,111,123,208],"rejected":[70],"otherwise.":[71],"This":[72,107],"criterion":[73,108,128,163,194],"fraction":[76],"variance":[78],"explained":[79],"($\\mathcal{V}^{ex}$),":[80],"measure":[82],"discrepancy":[85],"between":[86,206],"model":[88],"actual":[90],"samples.":[91],"$\\mathcal{V}^{ex}$":[92,162,173],"can":[93,195],"evaluate":[97],"impact":[99,186],"will":[102,184],"have":[103],"learning.":[106,188],"refines":[109],"improves":[112],"algorithm.":[116],"In":[117],"paper:":[119],"(1)":[120],"We":[121,136,170,190],"introduce":[122],"explore":[124],"$\\mathcal{V}^{ex}$,":[125],"selection":[127,180],"procedure.":[134],"(2)":[135],"conduct":[137],"experiments":[138],"across":[139],"variety":[141],"standard":[143],"benchmark":[144],"environments,":[145],"including":[146],"continuous":[147],"control":[148],"problems.":[149],"Our":[150],"results":[151],"show":[152,191],"better":[153],"performance":[154],"than":[155],"did":[158],"not":[159],"use":[160],"for":[164,178],"(3)":[169],"investigate":[171],"why":[172],"gives":[174],"good":[176],"evaluation":[177],"samples":[182],"that":[183],"positively":[185],"(4)":[189],"how":[192],"interpreted":[197],"as":[198],"dynamic":[200],"way":[201],"adjust":[203],"ratio":[205],"exploration":[207],"exploitation.":[209]},"counts_by_year":[{"year":2020,"cited_by_count":1},{"year":2019,"cited_by_count":1}],"updated_date":"2026-11-04T10:54:05.112534","created_date":"2019-04-11T00:00:00"}