{"id":"https://openalex.org/W7166891425","doi":"https://doi.org/10.18653/v1/2026.acl-long.363","title":"Do Not Step Into the Same River Twice: Learning to Reason from Trial and Error","display_name":"Do Not Step Into the Same River Twice: Learning to Reason from Trial and Error","publication_year":2026,"publication_date":"2026-01-01","ids":{"openalex":"https://openalex.org/W7166891425","doi":"https://doi.org/10.18653/v1/2026.acl-long.363"},"language":null,"primary_location":{"id":"doi:10.18653/v1/2026.acl-long.363","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.acl-long.363","pdf_url":"https://aclanthology.org/2026.acl-long.363.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2026.acl-long.363.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5139745583","display_name":"Chenming Tang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chenming Tang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139780451","display_name":"Hsiu-Yuan Huang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Hsiu-Yuan Huang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100668780","display_name":"Weijie Liu","orcid":"https://orcid.org/0000-0002-3054-766X"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Weijie Liu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5123014807","display_name":"Clive Bai","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Clive Bai","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5027918344","display_name":"Saiyong Yang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Saiyong Yang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5139771503","display_name":"Yunfang Wu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yunfang Wu","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":0.0,"has_fulltext":true,"cited_by_count":0,"citation_normalized_percentile":{"value":0.95541401,"is_in_top_1_percent":false,"is_in_top_10_percent":true},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"8018","last_page":"8032"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T13583","display_name":"Jury Decision Making Processes","score":0.04690000042319298,"subfield":{"id":"https://openalex.org/subfields/3308","display_name":"Law"},"field":{"id":"https://openalex.org/fields/33","display_name":"Social Sciences"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}},"topics":[{"id":"https://openalex.org/T13583","display_name":"Jury Decision Making Processes","score":0.04690000042319298,"subfield":{"id":"https://openalex.org/subfields/3308","display_name":"Law"},"field":{"id":"https://openalex.org/fields/33","display_name":"Social Sciences"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}},{"id":"https://openalex.org/T10318","display_name":"Criminal Justice and Corrections Analysis","score":0.023900000378489494,"subfield":{"id":"https://openalex.org/subfields/3312","display_name":"Sociology and Political Science"},"field":{"id":"https://openalex.org/fields/33","display_name":"Social Sciences"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}},{"id":"https://openalex.org/T10582","display_name":"Ethics in Clinical Research","score":0.02290000021457672,"subfield":{"id":"https://openalex.org/subfields/2739","display_name":"Public Health, Environmental and Occupational Health"},"field":{"id":"https://openalex.org/fields/27","display_name":"Medicine"},"domain":{"id":"https://openalex.org/domains/4","display_name":"Health Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/stability","display_name":"Stability (learning theory)","score":0.26170000433921814},{"id":"https://openalex.org/keywords/feature","display_name":"Feature (linguistics)","score":0.2587999999523163},{"id":"https://openalex.org/keywords/matching","display_name":"Matching (statistics)","score":0.25209999084472656},{"id":"https://openalex.org/keywords/set","display_name":"Set (abstract data type)","score":0.24279999732971191},{"id":"https://openalex.org/keywords/action","display_name":"Action (physics)","score":0.24130000174045563}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.4668999910354614},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3476000130176544},{"id":"https://openalex.org/C112972136","wikidata":"https://www.wikidata.org/wiki/Q7595718","display_name":"Stability (learning theory)","level":2,"score":0.26170000433921814},{"id":"https://openalex.org/C2776401178","wikidata":"https://www.wikidata.org/wiki/Q12050496","display_name":"Feature (linguistics)","level":2,"score":0.2587999999523163},{"id":"https://openalex.org/C165064840","wikidata":"https://www.wikidata.org/wiki/Q1321061","display_name":"Matching (statistics)","level":2,"score":0.25209999084472656},{"id":"https://openalex.org/C177264268","wikidata":"https://www.wikidata.org/wiki/Q1514741","display_name":"Set (abstract data type)","level":2,"score":0.24279999732971191},{"id":"https://openalex.org/C2780791683","wikidata":"https://www.wikidata.org/wiki/Q846785","display_name":"Action (physics)","level":2,"score":0.24130000174045563},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.2371000051498413},{"id":"https://openalex.org/C61797465","wikidata":"https://www.wikidata.org/wiki/Q1188986","display_name":"Term (time)","level":2,"score":0.23109999299049377},{"id":"https://openalex.org/C112930515","wikidata":"https://www.wikidata.org/wiki/Q4389547","display_name":"Risk analysis (engineering)","level":1,"score":0.2214999943971634}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2026.acl-long.363","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.acl-long.363","pdf_url":"https://aclanthology.org/2026.acl-long.363.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.18653/v1/2026.acl-long.363","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.acl-long.363","pdf_url":"https://aclanthology.org/2026.acl-long.363.pdf","source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 64th Annual Meeting of the Association for Computational Linguistics (Volume 1: Long Papers)","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[],"funders":[{"id":"https://openalex.org/F4320321001","display_name":"National Natural Science Foundation of China","ror":"https://ror.org/01h0zpd94"}],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7166891425.pdf","grobid_xml":"https://content.openalex.org/works/W7166891425.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Reinforcement":[0],"learning":[1],"with":[2,100],"verifiable":[3],"rewards":[4],"(RLVR)":[5],"has":[6],"significantly":[7],"boosted":[8],"the":[9,31,55,112,118],"reasoning":[10,138],"capability":[11,33],"of":[12,34,114],"language":[13],"models":[14],"(LMs).However,":[15],"existing":[16],"RLVR":[17],"approaches":[18,58],"train":[19],"LMs":[20,43,99],"based":[21],"on":[22,72,133],"their":[23,101],"own":[24],"on-policy":[25],"responses":[26],"and":[27,50,81,93,129,142,160,164],"are":[28],"constrained":[29],"by":[30,63,125],"initial":[32],"LMs,":[35],"thus":[36],"prone":[37],"to":[38,45,60,67,89],"exploration":[39,158,165],"stagnation,":[40],"in":[41,79,127,131],"which":[42,116],"fail":[44],"solve":[46],"more":[47],"training":[48,56,68],"problems":[49],"cannot":[51],"further":[52],"learn":[53],"from":[54,91],"data.Some":[57],"try":[59],"address":[61],"this":[62,83],"leveraging":[64],"off-policy":[65],"solutions":[66],"problems,":[69],"but":[70],"rely":[71],"external":[73,108,150],"expert":[74,109],"guidance":[75],"that":[76,97,148,154],"is":[77,169],"limited":[78],"availability":[80],"scalability.In":[82],"work,":[84],"we":[85],"propose":[86],"LTE":[87,155],"(Learning":[88],"reason":[90],"Trial":[92],"Error),":[94],"an":[95],"approach":[96],"hints":[98],"previously":[102],"self-made":[103],"mistakes,":[104],"not":[105],"requiring":[106],"any":[107],"guidance.Experiments":[110],"validate":[111],"effectiveness":[113],"LTE,":[115],"outperforms":[117],"normal":[119],"group":[120],"relative":[121],"policy":[122],"optimization":[123],"(GRPO)":[124],"5.02":[126],"Pass@1":[128],"9.96":[130],"Pass@k":[132],"average":[134],"across":[135],"six":[136],"mathematical":[137],"benchmarks":[139],"for":[140],"Qwen3-8B-Base":[141],"even":[143],"performs":[144],"better":[145],"than":[146],"methods":[147],"require":[149],"guidance.Further":[151],"analysis":[152],"confirms":[153],"successfully":[156],"mitigates":[157],"stagnation":[159],"enhances":[161],"both":[162],"exploitation":[163],"during":[166],"training.Our":[167],"code":[168],"available":[170],"at":[171],"https://github.com/":[172],"JamyDon/LTE.":[173]},"counts_by_year":[],"updated_date":"2026-07-29T14:22:42.915294","created_date":"2026-07-02T00:00:00"}