{"id":"https://openalex.org/W7155092078","doi":"https://doi.org/10.48550/arxiv.2604.18327","title":"PARM: Pipeline-Adapted Reward Model","display_name":"PARM: Pipeline-Adapted Reward Model","publication_year":2026,"publication_date":"2026-04-20","ids":{"openalex":"https://openalex.org/W7155092078","doi":"https://doi.org/10.48550/arxiv.2604.18327"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2604.18327","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.18327","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.2604.18327","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5134098127","display_name":"Xingyu Fan","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Fan, Xingyu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5134165710","display_name":"Wei Shao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Shao, Wei","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5134116052","display_name":"Jiacheng Liu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Liu, Jiacheng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5134139224","display_name":"Linqi Song","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Song, Linqi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5134205325","display_name":"Pheng Ann Heng","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Heng, Pheng Ann","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/T10028","display_name":"Topic Modeling","score":0.2985999882221222,"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/T10028","display_name":"Topic Modeling","score":0.2985999882221222,"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/T10181","display_name":"Natural Language Processing Techniques","score":0.08269999921321869,"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/T12535","display_name":"Machine Learning and Data Classification","score":0.04969999939203262,"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/pipeline","display_name":"Pipeline (software)","score":0.8044999837875366},{"id":"https://openalex.org/keywords/code","display_name":"Code (set theory)","score":0.582099974155426},{"id":"https://openalex.org/keywords/quality","display_name":"Quality (philosophy)","score":0.4697999954223633},{"id":"https://openalex.org/keywords/decoding-methods","display_name":"Decoding methods","score":0.451200008392334},{"id":"https://openalex.org/keywords/sampling","display_name":"Sampling (signal processing)","score":0.4016000032424927},{"id":"https://openalex.org/keywords/preference","display_name":"Preference","score":0.4000000059604645}],"concepts":[{"id":"https://openalex.org/C43521106","wikidata":"https://www.wikidata.org/wiki/Q2165493","display_name":"Pipeline (software)","level":2,"score":0.8044999837875366},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7376000285148621},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.582099974155426},{"id":"https://openalex.org/C2779530757","wikidata":"https://www.wikidata.org/wiki/Q1207505","display_name":"Quality (philosophy)","level":2,"score":0.4697999954223633},{"id":"https://openalex.org/C57273362","wikidata":"https://www.wikidata.org/wiki/Q576722","display_name":"Decoding methods","level":2,"score":0.451200008392334},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4462999999523163},{"id":"https://openalex.org/C140779682","wikidata":"https://www.wikidata.org/wiki/Q210868","display_name":"Sampling (signal processing)","level":3,"score":0.4016000032424927},{"id":"https://openalex.org/C2781249084","wikidata":"https://www.wikidata.org/wiki/Q908656","display_name":"Preference","level":2,"score":0.4000000059604645},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.39480000734329224},{"id":"https://openalex.org/C43126263","wikidata":"https://www.wikidata.org/wiki/Q128751","display_name":"Source code","level":2,"score":0.3555999994277954},{"id":"https://openalex.org/C2776207758","wikidata":"https://www.wikidata.org/wiki/Q5303302","display_name":"Downstream (manufacturing)","level":2,"score":0.3034000098705292},{"id":"https://openalex.org/C18762648","wikidata":"https://www.wikidata.org/wiki/Q42213","display_name":"Work (physics)","level":2,"score":0.30079999566078186},{"id":"https://openalex.org/C67186912","wikidata":"https://www.wikidata.org/wiki/Q367664","display_name":"Data modeling","level":2,"score":0.28700000047683716},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.27000001072883606},{"id":"https://openalex.org/C3019813237","wikidata":"https://www.wikidata.org/wiki/Q65089264","display_name":"Model validation","level":2,"score":0.26019999384880066},{"id":"https://openalex.org/C125411270","wikidata":"https://www.wikidata.org/wiki/Q18653","display_name":"Encoding (memory)","level":2,"score":0.25220000743865967}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2604.18327","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.18327","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.2604.18327","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2604.18327","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":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Reward":[0,85],"models":[1,9,56],"(RMs)":[2],"are":[3],"central":[4],"to":[5,96],"aligning":[6],"large":[7],"language":[8],"(LLMs)":[10],"with":[11,99],"human":[12],"preferences,":[13],"powering":[14],"RLHF":[15],"and":[16,60,73,92,113,124,129,149],"advanced":[17],"decoding":[18],"strategies.":[19],"While":[20],"most":[21],"prior":[22],"work":[23],"focuses":[24],"on":[25,116,136],"single-step":[26],"generation,":[27],"real-world":[28],"applications":[29],"increasingly":[30],"adopt":[31],"multi-stage":[32,158],"LLM":[33,159],"pipelines,":[34],"where":[35],"effective":[36],"reward":[37,55,70,155],"guidance":[38],"remains":[39],"underexplored.":[40],"We":[41,63,102],"investigate":[42],"this":[43],"through":[44],"code":[45,111],"generation":[46],"for":[47,157],"combinatorial":[48],"optimization,":[49],"constructing":[50],"a":[51,65,106],"pipeline":[52,75,108,146],"that":[53,142],"integrates":[54],"into":[57,154],"both":[58],"formulation":[59],"solution":[61],"stages.":[62],"identify":[64],"critical":[66],"challenge:":[67],"inconsistency":[68],"between":[69],"model":[71],"predictions":[72],"actual":[74],"execution":[76,122],"outcomes.":[77],"To":[78],"address":[79],"this,":[80],"we":[81],"propose":[82],"the":[83],"Pipeline-Adapted":[84],"Model":[86],"(PARM),":[87],"which":[88],"leverages":[89],"pipeline-specific":[90],"data":[91],"direct":[93],"preference":[94],"optimization":[95,119],"align":[97],"rewards":[98],"downstream":[100],"feedback.":[101],"instantiate":[103],"PARM":[104,143],"as":[105],"two-stage":[107],"(formulation":[109],"->":[110],"generation)":[112],"evaluate":[114],"it":[115],"four":[117],"public":[118],"benchmarks,":[120],"measuring":[121],"rate":[123],"solving":[125],"accuracy":[126],"against":[127],"baselines":[128],"sampling":[130],"methods.":[131],"A":[132],"supplementary":[133],"cross-domain":[134],"experiment":[135],"GSM8K":[137],"assesses":[138],"transferability.":[139],"Results":[140],"demonstrate":[141],"consistently":[144],"improves":[145],"output":[147],"quality":[148],"stability,":[150],"providing":[151],"new":[152],"insights":[153],"modeling":[156],"reasoning.":[160]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-04-22T00:00:00"}