{"id":"https://openalex.org/W7166658894","doi":"https://doi.org/10.48550/arxiv.2606.30537","title":"Learning from Mistakes: Rollout-Retrieval Lifelong Policy Learning for Autonomous Driving","display_name":"Learning from Mistakes: Rollout-Retrieval Lifelong Policy Learning for Autonomous Driving","publication_year":2026,"publication_date":"2026-06-29","ids":{"openalex":"https://openalex.org/W7166658894","doi":"https://doi.org/10.48550/arxiv.2606.30537"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2606.30537","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.30537","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2606.30537","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5139655096","display_name":"Cheng Gong","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Gong, Cheng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139685401","display_name":"Haoyang Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Haoyang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139656603","display_name":"Chao Lu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lu, Chao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139646916","display_name":"Zirui Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Zirui","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5100665612","display_name":"Jianwei Gong","orcid":"https://orcid.org/0000-0003-4651-8473"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Gong, Jianwei","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"cited_by_count":0,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.3037000000476837,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.3037000000476837,"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/T10462","display_name":"Reinforcement Learning in Robotics","score":0.2529999911785126,"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":0.1680999994277954,"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/bottleneck","display_name":"Bottleneck","score":0.7749999761581421},{"id":"https://openalex.org/keywords/lifelong-learning","display_name":"Lifelong learning","score":0.7451000213623047},{"id":"https://openalex.org/keywords/policy-learning","display_name":"Policy learning","score":0.6916000247001648},{"id":"https://openalex.org/keywords/software-deployment","display_name":"Software deployment","score":0.5674999952316284},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.5360000133514404},{"id":"https://openalex.org/keywords/generalization","display_name":"Generalization","score":0.5117999911308289},{"id":"https://openalex.org/keywords/planner","display_name":"Planner","score":0.4336000084877014}],"concepts":[{"id":"https://openalex.org/C2780513914","wikidata":"https://www.wikidata.org/wiki/Q18210350","display_name":"Bottleneck","level":2,"score":0.7749999761581421},{"id":"https://openalex.org/C108771440","wikidata":"https://www.wikidata.org/wiki/Q368475","display_name":"Lifelong learning","level":2,"score":0.7451000213623047},{"id":"https://openalex.org/C2779436431","wikidata":"https://www.wikidata.org/wiki/Q30672407","display_name":"Policy learning","level":2,"score":0.6916000247001648},{"id":"https://openalex.org/C105339364","wikidata":"https://www.wikidata.org/wiki/Q2297740","display_name":"Software deployment","level":2,"score":0.5674999952316284},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5526999831199646},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.5360000133514404},{"id":"https://openalex.org/C177148314","wikidata":"https://www.wikidata.org/wiki/Q170084","display_name":"Generalization","level":2,"score":0.5117999911308289},{"id":"https://openalex.org/C2776999362","wikidata":"https://www.wikidata.org/wiki/Q2349274","display_name":"Planner","level":2,"score":0.4336000084877014},{"id":"https://openalex.org/C2778137410","wikidata":"https://www.wikidata.org/wiki/Q2732820","display_name":"Government (linguistics)","level":2,"score":0.421099990606308},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4043000042438507},{"id":"https://openalex.org/C56739046","wikidata":"https://www.wikidata.org/wiki/Q192060","display_name":"Knowledge management","level":1,"score":0.3944999873638153},{"id":"https://openalex.org/C112930515","wikidata":"https://www.wikidata.org/wiki/Q4389547","display_name":"Risk analysis (engineering)","level":1,"score":0.38190001249313354},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.3465999960899353},{"id":"https://openalex.org/C13662910","wikidata":"https://www.wikidata.org/wiki/Q193139","display_name":"Trajectory","level":2,"score":0.31310001015663147},{"id":"https://openalex.org/C195094911","wikidata":"https://www.wikidata.org/wiki/Q14167904","display_name":"Process management","level":1,"score":0.29989999532699585},{"id":"https://openalex.org/C89611455","wikidata":"https://www.wikidata.org/wiki/Q6804646","display_name":"Mechanism (biology)","level":2,"score":0.28450000286102295},{"id":"https://openalex.org/C123587114","wikidata":"https://www.wikidata.org/wiki/Q2101508","display_name":"Policy analysis","level":2,"score":0.2806999981403351},{"id":"https://openalex.org/C2779305910","wikidata":"https://www.wikidata.org/wiki/Q5172809","display_name":"Corrective feedback","level":2,"score":0.275299996137619},{"id":"https://openalex.org/C109986646","wikidata":"https://www.wikidata.org/wiki/Q546113","display_name":"Public policy","level":2,"score":0.27129998803138733},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.2630000114440918}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2606.30537","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.30537","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2606.30537","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.30537","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I205783295","host_organization_name":"Cornell University","host_organization_lineage":["https://openalex.org/I205783295"],"host_organization_lineage_names":[],"type":"repository"},"license":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"sustainable_development_goals":[{"display_name":"Peace, Justice and strong institutions","id":"https://metadata.un.org/sdg/16","score":0.7202727198600769}],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Autonomous":[0],"driving":[1,60,87],"policies":[2,23],"should":[3,146],"be":[4],"able":[5],"to":[6,13,37,46,198],"improve":[7,72],"continually":[8,73],"as":[9],"deployment":[10],"exposes":[11],"them":[12],"increasingly":[14],"diverse":[15],"and":[16,31,48,112,153,169,186,209],"long-tail":[17,210],"traffic":[18],"situations.":[19],"However,":[20],"most":[21],"learning-based":[22,192],"are":[24],"trained":[25],"or":[26],"fine-tuned":[27],"on":[28,35,176,206],"expert":[29],"demonstrations":[30],"then":[32],"rely":[33],"largely":[34],"generalization":[36],"handle":[38],"challenging":[39,208],"closed-loop":[40,130,178,223],"scenarios,":[41],"lacking":[42],"an":[43],"explicit":[44],"mechanism":[45],"correct":[47],"retain":[49],"the":[50,114,134,144,202,207,216],"mistakes":[51,111,131,224],"exposed":[52],"in":[53,126,220],"these":[54],"scenarios.":[55],"This":[56],"paper":[57],"studies":[58],"autonomous":[59],"policy":[61,71,101,119,128,135,145,171,230],"improvement":[62],"from":[63,79,108],"a":[64,69,100,123,183,191],"lifelong":[65,118],"learning":[66,102],"perspective:":[67],"Can":[68],"pretrained":[70],"by":[74],"accumulating":[75],"corrective":[76,106,156,226],"knowledge":[77,116,166,227],"derived":[78],"its":[80],"own":[81],"mistakes,":[82],"while":[83],"retaining":[84],"previously":[85],"acquired":[86],"competence?":[88],"To":[89],"answer":[90],"this":[91],"question,":[92],"we":[93],"propose":[94],"Rollout-Retrieval":[95],"Lifelong":[96],"Policy":[97],"Learning":[98],"(R$^2$LPL),":[99],"framework":[103],"that":[104],"retrieves":[105],"targets":[107],"recoverable":[109,150,222],"policy-induced":[110],"retains":[113],"resulting":[115],"through":[117],"learning.":[120],"R^2LPL":[121],"addresses":[122],"key":[124],"bottleneck":[125],"continual":[127],"improvement:":[129],"reveal":[132],"where":[133],"is":[136],"weak,":[137],"but":[138],"do":[139],"not":[140],"directly":[141],"specify":[142],"what":[143],"learn.":[147],"By":[148],"filtering":[149],"mistake-related":[151],"states":[152],"retrieving":[154],"feasible":[155],"targets,":[157],"R$^2$LPL":[158,175,189,219],"turns":[159],"sparse":[160],"failure":[161],"evidence":[162],"into":[163,225],"compact":[164],"supervised":[165],"for":[167,228],"stable":[168],"sample-efficient":[170],"improvement.":[172,231],"We":[173],"evaluate":[174],"large-scale":[177],"nuPlan":[179],"benchmarks.":[180],"With":[181],"only":[182],"few":[184],"rollout":[185],"continual-learning":[187],"cycles,":[188],"elevates":[190],"planner":[193],"with":[194],"moderate":[195],"initial":[196],"performance":[197,200],"state-of-the-art":[199],"across":[201],"evaluated":[203],"benchmarks,":[204],"especially":[205],"Test14-hard":[211],"split.":[212],"These":[213],"results":[214],"demonstrate":[215],"effectiveness":[217],"of":[218],"converting":[221],"sustained":[229]},"counts_by_year":[],"updated_date":"2026-07-01T06:29:00.853634","created_date":"2026-07-01T00:00:00"}