{"id":"https://openalex.org/W7213611325","doi":"https://doi.org/10.48550/arxiv.2609.20756","title":"OPTED: On-Policy Fine-Tuning for End-to-End Driving using a Render-Free Teacher","display_name":"OPTED: On-Policy Fine-Tuning for End-to-End Driving using a Render-Free Teacher","publication_year":2026,"publication_date":"2026-09-17","ids":{"openalex":"https://openalex.org/W7213611325","doi":"https://doi.org/10.48550/arxiv.2609.20756"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2609.20756","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.20756","pdf_url":"https://arxiv.org/pdf/2609.20756","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"volume":null,"type":"preprint","indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2609.20756","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5152118618","display_name":"Damiano Da Col","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Da Col, Damiano","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5024136382","display_name":"Maximilian Igl","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Igl, Maximilian","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5069739104","display_name":"P\u00e9ter Karkus","orcid":"https://orcid.org/0000-0002-1474-9771","observed_orcids":["https://orcid.org/0000-0002-1474-9771"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Karkus, Peter","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5041189908","display_name":"Kashyap Chitta","orcid":"https://orcid.org/0000-0002-3891-3230","observed_orcids":["https://orcid.org/0000-0002-3891-3230"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chitta, Kashyap","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5091869385","display_name":"Boris Ivanovic","orcid":"https://orcid.org/0000-0002-8698-202X","observed_orcids":["https://orcid.org/0000-0002-8698-202X"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ivanovic, Boris","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5050003000","display_name":"Marco Pavone","orcid":"https://orcid.org/0000-0002-0206-4337","observed_orcids":["https://orcid.org/0000-0002-0206-4337"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Pavone, Marco","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5005404030","display_name":"Konrad Schindler","orcid":"https://orcid.org/0000-0002-3172-9246","observed_orcids":["https://orcid.org/0000-0002-3172-9246"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Schindler, Konrad","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5058481699","display_name":"Christos Sakaridis","orcid":"https://orcid.org/0000-0003-1127-8887","observed_orcids":["https://orcid.org/0000-0003-1127-8887"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Sakaridis, Christos","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":true,"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/T10462","display_name":"Reinforcement Learning in Robotics","score":0.8051999807357788,"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":0.8051999807357788,"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":0.1906999945640564,"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"}},{"id":"https://openalex.org/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.000699999975040555,"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/end-to-end-driving","display_name":"end-to-end driving","score":0.8119999766349792},{"id":"https://openalex.org/keywords/autonomous-driving","display_name":"autonomous driving","score":0.7829999923706055},{"id":"https://openalex.org/keywords/behavioral-cloning","display_name":"behavior cloning","score":0.6140000224113464},{"id":"https://openalex.org/keywords/human-demonstrations","display_name":"human demonstrations","score":0.36399999260902405},{"id":"https://openalex.org/keywords/distribution-shift","display_name":"distribution shift","score":0.5139999985694885}],"concepts":[{"id":"https://openalex.org/C105339364","wikidata":"https://www.wikidata.org/wiki/Q2297740","display_name":"Software deployment","level":2,"score":0.5889000296592712},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5489000082015991},{"id":"https://openalex.org/C97541855","wikidata":"https://www.wikidata.org/wiki/Q830687","display_name":"Reinforcement learning","level":2,"score":0.5019000172615051},{"id":"https://openalex.org/C63584917","wikidata":"https://www.wikidata.org/wiki/Q333286","display_name":"Bounding overwatch","level":2,"score":0.47699999809265137},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.43220001459121704},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.41110000014305115},{"id":"https://openalex.org/C2780626000","wikidata":"https://www.wikidata.org/wiki/Q5936775","display_name":"Human-in-the-loop","level":2,"score":0.3824999928474426},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.34769999980926514},{"id":"https://openalex.org/C44154836","wikidata":"https://www.wikidata.org/wiki/Q45045","display_name":"Simulation","level":1,"score":0.32510000467300415},{"id":"https://openalex.org/C133462117","wikidata":"https://www.wikidata.org/wiki/Q4929239","display_name":"Data collection","level":2,"score":0.3206000030040741},{"id":"https://openalex.org/C207673951","wikidata":"https://www.wikidata.org/wiki/Q1303150","display_name":"Compounding","level":2,"score":0.29840001463890076},{"id":"https://openalex.org/C107457646","wikidata":"https://www.wikidata.org/wiki/Q207434","display_name":"Human\u2013computer interaction","level":1,"score":0.2980000078678131},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.29190000891685486},{"id":"https://openalex.org/C2780689630","wikidata":"https://www.wikidata.org/wiki/Q2081815","display_name":"Driving simulator","level":2,"score":0.289000004529953},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.2727000117301941},{"id":"https://openalex.org/C67203356","wikidata":"https://www.wikidata.org/wiki/Q1321905","display_name":"Reinforcement","level":2,"score":0.2583000063896179},{"id":"https://openalex.org/C2776401178","wikidata":"https://www.wikidata.org/wiki/Q12050496","display_name":"Feature (linguistics)","level":2,"score":0.25450000166893005},{"id":"https://openalex.org/C117896860","wikidata":"https://www.wikidata.org/wiki/Q11376","display_name":"Acceleration","level":2,"score":0.2526000142097473},{"id":"https://openalex.org/C127413603","wikidata":"https://www.wikidata.org/wiki/Q11023","display_name":"Engineering","level":0,"score":0.25049999356269836}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2609.20756","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.20756","pdf_url":"https://arxiv.org/pdf/2609.20756","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2609.20756","doi":"https://doi.org/10.48550/arxiv.2609.20756","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2609.20756","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"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"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2609.20756","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.20756","pdf_url":"https://arxiv.org/pdf/2609.20756","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"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":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7213611325.pdf","grobid_xml":"https://content.openalex.org/works/W7213611325.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":["https://openalex.org/W7165010320","https://openalex.org/W7117078700","https://openalex.org/W7167619160","https://openalex.org/W7160547054","https://openalex.org/W4416969221","https://openalex.org/W7214892509","https://openalex.org/W7138073728","https://openalex.org/W7138382548","https://openalex.org/W7138329672","https://openalex.org/W7214658492"],"abstract_inverted_index":{"As":[0],"scaling":[1],"pre-training":[2],"data":[3,48],"alone":[4],"yields":[5],"diminishing":[6],"returns,":[7],"post-training":[8,57,83],"is":[9,91],"becoming":[10],"increasingly":[11],"important":[12],"across":[13],"physical":[14],"AI":[15],"domains":[16],"such":[17],"as":[18],"autonomous":[19],"driving.":[20],"End-to-end":[21],"driving":[22,135],"policies":[23],"are":[24],"pre-trained":[25,109],"in":[26,127],"open":[27],"loop":[28],"with":[29,154],"behavior":[30],"cloning":[31],"on":[32,95],"human":[33,172],"demonstrations.":[34],"However,":[35],"compounding":[36],"errors":[37],"during":[38,111],"closed-loop":[39,112,152],"deployment":[40],"can":[41,58],"take":[42],"the":[43,46,51,82,85,108,171],"vehicle":[44],"outside":[45],"training":[47],"distribution,":[49],"increasing":[50],"risk":[52,61],"of":[53,84,133,142,158],"safety-critical":[54],"incidents.":[55],"Closed-loop":[56],"mitigate":[59],"this":[60],"but":[62],"requires":[63],"costly":[64],"simulation":[65],"for":[66,74],"sensor-based":[67],"policies.":[68],"We":[69,114],"propose":[70],"OPTED":[71,116,150],"(on-policy":[72],"fine-tuning":[73],"end-to-end":[75,86],"driving)":[76],"which":[77],"decouples":[78],"reinforcement":[79],"learning":[80],"from":[81],"policy:":[87],"a":[88],"privileged":[89],"teacher":[90,103],"trained":[92],"using":[93,129],"RL":[94,165],"vectorized":[96],"inputs":[97],"(HD-map":[98],"and":[99,122,124,144],"bounding":[100],"boxes).":[101],"This":[102],"then":[104],"provides":[105],"supervision":[106],"to":[107,117,170],"student":[110],"post-training.":[113],"apply":[115],"two":[118],"camera-based":[119],"models,":[120],"TransFuser":[121],"VaVAM,":[123],"fine-tune":[125],"them":[126],"AlpaSim,":[128],"neural":[130],"reconstructions":[131],"(3DGS)":[132],"real":[134],"logs.":[136],"Driving":[137],"scores":[138],"increase":[139],"by":[140],"factors":[141],"1.6$\\times$":[143],"9.5$\\times$,":[145],"respectively.":[146],"In":[147],"controlled":[148],"experiments":[149],"matches":[151],"performance":[153],"approximately":[155],"three":[156],"orders":[157],"magnitude":[159],"fewer":[160],"simulator":[161],"interactions":[162],"than":[163],"direct":[164],"post-training,":[166],"while":[167],"staying":[168],"closer":[169],"prior.":[173],"Project":[174],"page:":[175],"https://01dami23.github.io/opted/":[176]},"counts_by_year":[],"updated_date":"2026-10-28T10:33:24.094632","created_date":"2026-09-19T00:00:00"}