{"id":"https://openalex.org/W4306167947","doi":"https://doi.org/10.1109/tpami.2022.3213503","title":"Robust Losses for Learning Value Functions","display_name":"Robust Losses for Learning Value Functions","publication_year":2022,"publication_date":"2022-01-01","ids":{"openalex":"https://openalex.org/W4306167947","doi":"https://doi.org/10.1109/tpami.2022.3213503","pmid":"https://pubmed.ncbi.nlm.nih.gov/36227822"},"language":"en","primary_location":{"id":"doi:10.1109/tpami.2022.3213503","is_oa":true,"landing_page_url":"https://doi.org/10.1109/tpami.2022.3213503","pdf_url":"https://ieeexplore.ieee.org/ielx7/34/4359286/09918637.pdf","source":{"id":"https://openalex.org/S199944782","display_name":"IEEE Transactions on Pattern Analysis and Machine Intelligence","issn_l":"0162-8828","issn":["0162-8828","1939-3539","2160-9292"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310320439","host_organization_name":"IEEE Computer Society","host_organization_lineage":["https://openalex.org/P4310320439","https://openalex.org/P4310319808"],"host_organization_lineage_names":["IEEE Computer Society","Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Transactions on Pattern Analysis and Machine Intelligence","raw_type":"journal-article"},"type":"article","indexed_in":["crossref","pubmed"],"open_access":{"is_oa":true,"oa_status":"hybrid","oa_url":"https://ieeexplore.ieee.org/ielx7/34/4359286/09918637.pdf","any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5103118703","display_name":"Andrew Patterson","orcid":"https://orcid.org/0000-0003-2461-3103"},"institutions":[{"id":"https://openalex.org/I154425047","display_name":"University of Alberta","ror":"https://ror.org/0160cpw27","country_code":"CA","type":"education","lineage":["https://openalex.org/I154425047"]}],"countries":["CA"],"is_corresponding":false,"raw_author_name":"Andrew Patterson","raw_affiliation_strings":["Computing Science, University of Alberta, Edmonton, Canada"],"raw_orcid":"https://orcid.org/0000-0003-2461-3103","affiliations":[{"raw_affiliation_string":"Computing Science, University of Alberta, Edmonton, Canada","institution_ids":["https://openalex.org/I154425047"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5042936131","display_name":"Victor Liao","orcid":"https://orcid.org/0000-0001-5998-7944"},"institutions":[{"id":"https://openalex.org/I151746483","display_name":"University of Waterloo","ror":"https://ror.org/01aff2v68","country_code":"CA","type":"education","lineage":["https://openalex.org/I151746483"]}],"countries":["CA"],"is_corresponding":false,"raw_author_name":"Victor Liao","raw_affiliation_strings":["Faculty of Mathematics, University of Waterloo, Waterloo, Canada"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Faculty of Mathematics, University of Waterloo, Waterloo, Canada","institution_ids":["https://openalex.org/I151746483"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5101613484","display_name":"Martha White","orcid":"https://orcid.org/0000-0002-5356-2950"},"institutions":[{"id":"https://openalex.org/I154425047","display_name":"University of Alberta","ror":"https://ror.org/0160cpw27","country_code":"CA","type":"education","lineage":["https://openalex.org/I154425047"]}],"countries":["CA"],"is_corresponding":false,"raw_author_name":"Martha White","raw_affiliation_strings":["Computing Science, University of Alberta, Edmonton, Canada"],"raw_orcid":"https://orcid.org/0000-0002-5356-2950","affiliations":[{"raw_affiliation_string":"Computing Science, University of Alberta, Edmonton, Canada","institution_ids":["https://openalex.org/I154425047"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":3,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":2.081,"has_fulltext":true,"cited_by_count":16,"citation_normalized_percentile":{"value":0.88943004,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":{"min":90,"max":99},"biblio":{"volume":"45","issue":"5","first_page":"1","last_page":"12"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10462","display_name":"Reinforcement Learning in Robotics","score":0.9998999834060669,"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":0.9998999834060669,"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/T12101","display_name":"Advanced Bandit Algorithms Research","score":0.9927999973297119,"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"}},{"id":"https://openalex.org/T11612","display_name":"Stochastic Gradient Optimization Techniques","score":0.9921000003814697,"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"}}],"keywords":[{"id":"https://openalex.org/keywords/reinforcement-learning","display_name":"Reinforcement learning","score":0.7627837061882019},{"id":"https://openalex.org/keywords/mean-squared-error","display_name":"Mean squared error","score":0.7333437204360962},{"id":"https://openalex.org/keywords/outlier","display_name":"Outlier","score":0.6951996088027954},{"id":"https://openalex.org/keywords/clipping","display_name":"Clipping (morphology)","score":0.6486232876777649},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.574074387550354},{"id":"https://openalex.org/keywords/bellman-equation","display_name":"Bellman equation","score":0.5315002799034119},{"id":"https://openalex.org/keywords/variance","display_name":"Variance (accounting)","score":0.4695371389389038},{"id":"https://openalex.org/keywords/mathematical-optimization","display_name":"Mathematical optimization","score":0.4567207098007202},{"id":"https://openalex.org/keywords/robust-statistics","display_name":"Robust statistics","score":0.4503912627696991},{"id":"https://openalex.org/keywords/function","display_name":"Function (biology)","score":0.43404263257980347},{"id":"https://openalex.org/keywords/robust-control","display_name":"Robust control","score":0.43148359656333923},{"id":"https://openalex.org/keywords/sensitivity","display_name":"Sensitivity (control systems)","score":0.42432790994644165},{"id":"https://openalex.org/keywords/robust-regression","display_name":"Robust regression","score":0.41860151290893555},{"id":"https://openalex.org/keywords/robustness","display_name":"Robustness (evolution)","score":0.4154735803604126},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.33591610193252563},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.2543427348136902},{"id":"https://openalex.org/keywords/statistics","display_name":"Statistics","score":0.21890410780906677},{"id":"https://openalex.org/keywords/control-system","display_name":"Control system","score":0.15931817889213562}],"concepts":[{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.7627837061882019},{"id":"https://openalex.org/C139945424","wikidata":"https://www.wikidata.org/wiki/Q1940696","display_name":"Mean squared error","level":2,"score":0.7333437204360962},{"id":"https://openalex.org/C79337645","wikidata":"https://www.wikidata.org/wiki/Q779824","display_name":"Outlier","level":2,"score":0.6951996088027954},{"id":"https://openalex.org/C2776848632","wikidata":"https://www.wikidata.org/wiki/Q853463","display_name":"Clipping (morphology)","level":2,"score":0.6486232876777649},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.574074387550354},{"id":"https://openalex.org/C14646407","wikidata":"https://www.wikidata.org/wiki/Q1430750","display_name":"Bellman equation","level":2,"score":0.5315002799034119},{"id":"https://openalex.org/C196083921","wikidata":"https://www.wikidata.org/wiki/Q7915758","display_name":"Variance (accounting)","level":2,"score":0.4695371389389038},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.4567207098007202},{"id":"https://openalex.org/C67226441","wikidata":"https://www.wikidata.org/wiki/Q1665389","display_name":"Robust statistics","level":3,"score":0.4503912627696991},{"id":"https://openalex.org/C14036430","wikidata":"https://www.wikidata.org/wiki/Q3736076","display_name":"Function (biology)","level":2,"score":0.43404263257980347},{"id":"https://openalex.org/C31531917","wikidata":"https://www.wikidata.org/wiki/Q915157","display_name":"Robust control","level":3,"score":0.43148359656333923},{"id":"https://openalex.org/C21200559","wikidata":"https://www.wikidata.org/wiki/Q7451068","display_name":"Sensitivity (control systems)","level":2,"score":0.42432790994644165},{"id":"https://openalex.org/C70259352","wikidata":"https://www.wikidata.org/wiki/Q1847839","display_name":"Robust regression","level":3,"score":0.41860151290893555},{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.4154735803604126},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.33591610193252563},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.2543427348136902},{"id":"https://openalex.org/C105795698","wikidata":"https://www.wikidata.org/wiki/Q12483","display_name":"Statistics","level":1,"score":0.21890410780906677},{"id":"https://openalex.org/C17500928","wikidata":"https://www.wikidata.org/wiki/Q959968","display_name":"Control system","level":2,"score":0.15931817889213562},{"id":"https://openalex.org/C185592680","wikidata":"https://www.wikidata.org/wiki/Q2329","display_name":"Chemistry","level":0,"score":0.0},{"id":"https://openalex.org/C55493867","wikidata":"https://www.wikidata.org/wiki/Q7094","display_name":"Biochemistry","level":1,"score":0.0},{"id":"https://openalex.org/C121955636","wikidata":"https://www.wikidata.org/wiki/Q4116214","display_name":"Accounting","level":1,"score":0.0},{"id":"https://openalex.org/C104317684","wikidata":"https://www.wikidata.org/wiki/Q7187","display_name":"Gene","level":2,"score":0.0},{"id":"https://openalex.org/C24326235","wikidata":"https://www.wikidata.org/wiki/Q126095","display_name":"Electronic engineering","level":1,"score":0.0},{"id":"https://openalex.org/C119599485","wikidata":"https://www.wikidata.org/wiki/Q43035","display_name":"Electrical engineering","level":1,"score":0.0},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.0},{"id":"https://openalex.org/C41895202","wikidata":"https://www.wikidata.org/wiki/Q8162","display_name":"Linguistics","level":1,"score":0.0},{"id":"https://openalex.org/C78458016","wikidata":"https://www.wikidata.org/wiki/Q840400","display_name":"Evolutionary biology","level":1,"score":0.0},{"id":"https://openalex.org/C86803240","wikidata":"https://www.wikidata.org/wiki/Q420","display_name":"Biology","level":0,"score":0.0},{"id":"https://openalex.org/C144133560","wikidata":"https://www.wikidata.org/wiki/Q4830453","display_name":"Business","level":0,"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":2,"locations":[{"id":"doi:10.1109/tpami.2022.3213503","is_oa":true,"landing_page_url":"https://doi.org/10.1109/tpami.2022.3213503","pdf_url":"https://ieeexplore.ieee.org/ielx7/34/4359286/09918637.pdf","source":{"id":"https://openalex.org/S199944782","display_name":"IEEE Transactions on Pattern Analysis and Machine Intelligence","issn_l":"0162-8828","issn":["0162-8828","1939-3539","2160-9292"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310320439","host_organization_name":"IEEE Computer Society","host_organization_lineage":["https://openalex.org/P4310320439","https://openalex.org/P4310319808"],"host_organization_lineage_names":["IEEE Computer Society","Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Transactions on Pattern Analysis and Machine Intelligence","raw_type":"journal-article"},{"id":"pmid:36227822","is_oa":false,"landing_page_url":"https://pubmed.ncbi.nlm.nih.gov/36227822","pdf_url":null,"source":{"id":"https://openalex.org/S4306525036","display_name":"PubMed","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I1299303238","host_organization_name":"National Institutes of Health","host_organization_lineage":["https://openalex.org/I1299303238"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE transactions on pattern analysis and machine intelligence","raw_type":null}],"best_oa_location":{"id":"doi:10.1109/tpami.2022.3213503","is_oa":true,"landing_page_url":"https://doi.org/10.1109/tpami.2022.3213503","pdf_url":"https://ieeexplore.ieee.org/ielx7/34/4359286/09918637.pdf","source":{"id":"https://openalex.org/S199944782","display_name":"IEEE Transactions on Pattern Analysis and Machine Intelligence","issn_l":"0162-8828","issn":["0162-8828","1939-3539","2160-9292"],"is_oa":false,"is_in_doaj":false,"is_core":true,"host_organization":"https://openalex.org/P4310320439","host_organization_name":"IEEE Computer Society","host_organization_lineage":["https://openalex.org/P4310320439","https://openalex.org/P4310319808"],"host_organization_lineage_names":["IEEE Computer Society","Institute of Electrical and Electronics Engineers"],"type":"journal"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"IEEE Transactions on Pattern Analysis and Machine Intelligence","raw_type":"journal-article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W4306167947.pdf","grobid_xml":"https://content.openalex.org/works/W4306167947.grobid-xml"},"referenced_works_count":38,"referenced_works":["https://openalex.org/W41554520","https://openalex.org/W1646707810","https://openalex.org/W1992208280","https://openalex.org/W2075268401","https://openalex.org/W2084493382","https://openalex.org/W2104753538","https://openalex.org/W2139418546","https://openalex.org/W2150459019","https://openalex.org/W2294586855","https://openalex.org/W2746553466","https://openalex.org/W2761873684","https://openalex.org/W2765302304","https://openalex.org/W2891076394","https://openalex.org/W2902098903","https://openalex.org/W2948937890","https://openalex.org/W3041202696","https://openalex.org/W3216772467","https://openalex.org/W4238862570","https://openalex.org/W4287197771","https://openalex.org/W4298857966","https://openalex.org/W6612245137","https://openalex.org/W6637967152","https://openalex.org/W6682142764","https://openalex.org/W6683300800","https://openalex.org/W6685444567","https://openalex.org/W6697391007","https://openalex.org/W6725941776","https://openalex.org/W6732811253","https://openalex.org/W6752164047","https://openalex.org/W6762719077","https://openalex.org/W6763056091","https://openalex.org/W6779199806","https://openalex.org/W6779390608","https://openalex.org/W6780559895","https://openalex.org/W6787020633","https://openalex.org/W6790675985","https://openalex.org/W6794007783","https://openalex.org/W6804601995"],"related_works":["https://openalex.org/W4210813465","https://openalex.org/W2170905369","https://openalex.org/W2135642348","https://openalex.org/W2151976295","https://openalex.org/W2886594279","https://openalex.org/W2922716635","https://openalex.org/W3159798879","https://openalex.org/W2521370002","https://openalex.org/W1990929462","https://openalex.org/W4306167947"],"abstract_inverted_index":{"Most":[0],"value":[1],"function":[2],"learning":[3,7],"algorithms":[4,178],"in":[5,36,48,134],"reinforcement":[6],"are":[8,20,73,179],"based":[9],"on":[10,75,91],"the":[11,29,32,70,136,145,148,154,158,166,175],"mean":[12,167],"squared":[13,18,95,168],"(projected)":[14],"Bellman":[15,96,111,115,169],"error.":[16,116,170],"However,":[17],"errors":[19,97],"known":[21,84],"to":[22,25,64,67,130,190],"be":[23,65],"sensitive":[24],"outliers,":[26],"both":[27,135,183],"skewing":[28],"solution":[30],"of":[31,122,147],"objective":[33],"and":[34,38,103,113,140,185],"resulting":[35,176],"high-magnitude":[37,44],"high-variance":[39],"gradients.":[40],"To":[41],"control":[42,141],"these":[43,61,132],"updates,":[45],"typical":[46],"strategies":[47,62],"RL":[49],"involve":[50],"clipping":[51,53,58],"gradients,":[52],"rewards,":[54,56],"rescaling":[55],"or":[57],"errors.":[59],"While":[60],"appear":[63],"related":[66],"robust":[68,123,149,159],"losses-like":[69],"Huber":[71,110],"loss-they":[72],"built":[74],"semi-gradient":[76],"update":[77],"rules":[78],"which":[79],"do":[80],"not":[81],"minimize":[82,131],"a":[83,99,105,109,120],"loss.":[85],"In":[86],"this":[87],"work,":[88],"we":[89,172],"build":[90],"recent":[92],"insights":[93],"reformulating":[94],"as":[98],"saddlepoint":[100,106],"optimization":[101],"problem":[102,155],"propose":[104],"reformulation":[107],"for":[108,182],"error":[112],"Absolute":[114],"We":[117,143],"start":[118],"from":[119],"formalization":[121],"losses,":[124,150],"then":[125],"derive":[126],"sound":[127],"gradient-based":[128,177],"approaches":[129],"losses":[133,160],"online":[137],"off-policy":[138],"prediction":[139,184],"settings.":[142],"characterize":[144],"solutions":[146,164],"providing":[151],"insight":[152],"into":[153],"settings":[156],"where":[157],"define":[161],"notably":[162],"better":[163],"than":[165],"Finally,":[171],"show":[173],"that":[174],"more":[180],"stable,":[181],"control,":[186],"with":[187],"less":[188],"sensitivity":[189],"meta-parameters.":[191]},"counts_by_year":[{"year":2026,"cited_by_count":1},{"year":2025,"cited_by_count":6},{"year":2024,"cited_by_count":8},{"year":2023,"cited_by_count":1}],"updated_date":"2026-06-11T09:08:48.828518","created_date":"2025-10-10T00:00:00"}