{"id":"https://openalex.org/W7163373499","doi":"https://doi.org/10.48550/arxiv.2606.02615","title":"FSA-GRPO: Teaching Auditory LLMs to Use Few-shot Demonstrations","display_name":"FSA-GRPO: Teaching Auditory LLMs to Use Few-shot Demonstrations","publication_year":2026,"publication_date":"2026-05-26","ids":{"openalex":"https://openalex.org/W7163373499","doi":"https://doi.org/10.48550/arxiv.2606.02615"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2606.02615","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.02615","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.2606.02615","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5137749335","display_name":"Haolong Zheng","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zheng, Haolong","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5017974335","display_name":"Siyin Wang","orcid":"https://orcid.org/0009-0008-7992-2476"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wang, Siyin","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5137786477","display_name":"Xulin Fan","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Fan, Xulin","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5137785093","display_name":"Zengrui Jin","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jin, Zengrui","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5137757947","display_name":"Mark Hasegawa-Johnson","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Hasegawa-Johnson, Mark","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/T10201","display_name":"Speech Recognition and Synthesis","score":0.4726000130176544,"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/T10201","display_name":"Speech Recognition and Synthesis","score":0.4726000130176544,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.3352999985218048,"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/T10028","display_name":"Topic Modeling","score":0.028300000354647636,"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/leverage","display_name":"Leverage (statistics)","score":0.6453999876976013},{"id":"https://openalex.org/keywords/weighting","display_name":"Weighting","score":0.5375999808311462},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.5306000113487244},{"id":"https://openalex.org/keywords/limiting","display_name":"Limiting","score":0.5044000148773193},{"id":"https://openalex.org/keywords/adaptation","display_name":"Adaptation (eye)","score":0.4729999899864197},{"id":"https://openalex.org/keywords/selection","display_name":"Selection (genetic algorithm)","score":0.4431000053882599},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.4404999911785126}],"concepts":[{"id":"https://openalex.org/C153083717","wikidata":"https://www.wikidata.org/wiki/Q6535263","display_name":"Leverage (statistics)","level":2,"score":0.6453999876976013},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6313999891281128},{"id":"https://openalex.org/C183115368","wikidata":"https://www.wikidata.org/wiki/Q856577","display_name":"Weighting","level":2,"score":0.5375999808311462},{"id":"https://openalex.org/C28490314","wikidata":"https://www.wikidata.org/wiki/Q189436","display_name":"Speech recognition","level":1,"score":0.531499981880188},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.5306000113487244},{"id":"https://openalex.org/C188198153","wikidata":"https://www.wikidata.org/wiki/Q1613840","display_name":"Limiting","level":2,"score":0.5044000148773193},{"id":"https://openalex.org/C139807058","wikidata":"https://www.wikidata.org/wiki/Q352374","display_name":"Adaptation (eye)","level":2,"score":0.4729999899864197},{"id":"https://openalex.org/C81917197","wikidata":"https://www.wikidata.org/wiki/Q628760","display_name":"Selection (genetic algorithm)","level":2,"score":0.4431000053882599},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.4404999911785126},{"id":"https://openalex.org/C15744967","wikidata":"https://www.wikidata.org/wiki/Q9418","display_name":"Psychology","level":0,"score":0.38670000433921814},{"id":"https://openalex.org/C180747234","wikidata":"https://www.wikidata.org/wiki/Q23373","display_name":"Cognitive psychology","level":1,"score":0.36250001192092896},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.3452000021934509},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.34389999508857727},{"id":"https://openalex.org/C74672266","wikidata":"https://www.wikidata.org/wiki/Q815859","display_name":"Language acquisition","level":2,"score":0.34139999747276306},{"id":"https://openalex.org/C44725695","wikidata":"https://www.wikidata.org/wiki/Q288156","display_name":"Normative","level":2,"score":0.3310000002384186},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.32679998874664307},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3249000012874603},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.3095000088214874},{"id":"https://openalex.org/C155635449","wikidata":"https://www.wikidata.org/wiki/Q4674699","display_name":"Acoustic model","level":3,"score":0.27410000562667847},{"id":"https://openalex.org/C171041071","wikidata":"https://www.wikidata.org/wiki/Q36870","display_name":"First language","level":2,"score":0.2685999870300293},{"id":"https://openalex.org/C61328038","wikidata":"https://www.wikidata.org/wiki/Q3358061","display_name":"Speech processing","level":2,"score":0.25999999046325684},{"id":"https://openalex.org/C2776230583","wikidata":"https://www.wikidata.org/wiki/Q1322198","display_name":"Spoken language","level":2,"score":0.25099998712539673}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2606.02615","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.02615","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.2606.02615","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.02615","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":[{"score":0.40214231610298157,"id":"https://metadata.un.org/sdg/4","display_name":"Quality Education"}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Few-shot":[0],"prompting":[1],"provides":[2],"an":[3,58,124],"effective":[4,125,146],"way":[5],"to":[6,12,30,40,68,72,122],"adapt":[7],"auditory":[8,22],"large":[9,23],"language":[10,24],"models":[11,25],"low-resource":[13],"tasks":[14],"such":[15],"as":[16],"children's":[17,102],"speech":[18,103,108],"recognition.":[19],"However,":[20],"most":[21],"are":[26,135],"not":[27,99],"explicitly":[28],"trained":[29],"perform":[31],"inference":[32],"in":[33,101,107],"this":[34,50],"demonstration-conditioned":[35],"format,":[36],"limiting":[37],"the":[38,70,91],"extent":[39],"which":[41],"they":[42],"can":[43],"benefit":[44],"from":[45],"few-shot":[46,74,79,94],"prompting.":[47],"To":[48],"address":[49],"limitation,":[51],"we":[52],"introduce":[53],"Few-Shot":[54],"Aware":[55],"GRPO":[56],"(FSA-GRPO),":[57],"RL-based":[59],"post-training":[60],"recipe":[61],"that":[62,131],"uses":[63],"a":[64],"specially":[65],"designed":[66],"reward":[67,120],"encourage":[69],"model":[71],"leverage":[73],"demonstrations,":[75],"thereby":[76],"strengthening":[77],"its":[78],"adaptation":[80,95],"ability.":[81],"Notably,":[82],"training":[83,126],"with":[84],"only":[85,100],"high-resource":[86],"adult":[87],"ASR":[88],"data":[89,116,134],"improves":[90],"model's":[92],"general":[93],"ability,":[96],"yielding":[97],"gains":[98],"recognition":[104],"but":[105],"also":[106],"translation":[109],"and":[110,118],"audio":[111],"understanding.":[112],"We":[113],"further":[114],"study":[115],"selection":[117],"auxiliary":[119],"weighting":[121],"identify":[123],"recipe.":[127],"Our":[128],"experiments":[129],"show":[130],"when":[132],"in-domain":[133],"unavailable":[136],"or":[137],"cannot":[138],"be":[139],"used":[140],"for":[141],"training,":[142],"FSA-GRPO":[143],"is":[144],"more":[145],"than":[147],"direct":[148],"tuning":[149],"on":[150],"related":[151],"out-of-domain":[152],"data.":[153]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-06-04T00:00:00"}