{"id":"https://openalex.org/W4412377085","doi":"https://doi.org/10.1145/3726302.3730063","title":"Preference-Strength-Aware Self-Improving Alignment with Generative Preference Models","display_name":"Preference-Strength-Aware Self-Improving Alignment with Generative Preference Models","publication_year":2025,"publication_date":"2025-07-13","ids":{"openalex":"https://openalex.org/W4412377085","doi":"https://doi.org/10.1145/3726302.3730063"},"language":"en","primary_location":{"id":"doi:10.1145/3726302.3730063","is_oa":false,"landing_page_url":"https://doi.org/10.1145/3726302.3730063","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 48th International ACM SIGIR Conference on Research and Development in Information Retrieval","raw_type":"proceedings-article"},"type":"article","indexed_in":["crossref"],"open_access":{"is_oa":false,"oa_status":"closed","oa_url":null,"any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5073132517","display_name":"Yuanzhao Zhai","orcid":"https://orcid.org/0000-0003-1385-0074"},"institutions":[{"id":"https://openalex.org/I170215575","display_name":"National University of Defense Technology","ror":"https://ror.org/05d2yfz11","country_code":"CN","type":"education","lineage":["https://openalex.org/I170215575"]}],"countries":["CN"],"is_corresponding":true,"raw_author_name":"Yuanzhao Zhai","raw_affiliation_strings":["National University of Defense Technology, Changsha, China and State Key Laboratory of Complex & Critical Software Environment, Changsha, China"],"affiliations":[{"raw_affiliation_string":"National University of Defense Technology, Changsha, China and State Key Laboratory of Complex & Critical Software Environment, Changsha, China","institution_ids":["https://openalex.org/I170215575"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5108148866","display_name":"Zhuo Zhang","orcid":"https://orcid.org/0000-0002-1835-5411"},"institutions":[{"id":"https://openalex.org/I204983213","display_name":"Harbin Institute of Technology","ror":"https://ror.org/01yqg2h08","country_code":"CN","type":"education","lineage":["https://openalex.org/I204983213"]},{"id":"https://openalex.org/I4210136793","display_name":"Peng Cheng Laboratory","ror":"https://ror.org/03qdqbt06","country_code":"CN","type":"facility","lineage":["https://openalex.org/I4210136793"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zhuo Zhang","raw_affiliation_strings":["Harbin Institute of Technology (Shenzhen), Shenzhen, China and Peng Cheng Laboratory, Shenzhen, China"],"affiliations":[{"raw_affiliation_string":"Harbin Institute of Technology (Shenzhen), Shenzhen, China and Peng Cheng Laboratory, Shenzhen, China","institution_ids":["https://openalex.org/I4210136793","https://openalex.org/I204983213"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5078272955","display_name":"Cheng Yang","orcid":"https://orcid.org/0000-0002-4782-1645"},"institutions":[{"id":"https://openalex.org/I170215575","display_name":"National University of Defense Technology","ror":"https://ror.org/05d2yfz11","country_code":"CN","type":"education","lineage":["https://openalex.org/I170215575"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Cheng Yang","raw_affiliation_strings":["National University of Defense Technology, Changsha, Hunan, China and State Key Laboratory of Complex & Critical Software Environment, Changsha, China"],"affiliations":[{"raw_affiliation_string":"National University of Defense Technology, Changsha, Hunan, China and State Key Laboratory of Complex & Critical Software Environment, Changsha, China","institution_ids":["https://openalex.org/I170215575"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5013340793","display_name":"Kele Xu","orcid":"https://orcid.org/0000-0001-5997-5169"},"institutions":[{"id":"https://openalex.org/I170215575","display_name":"National University of Defense Technology","ror":"https://ror.org/05d2yfz11","country_code":"CN","type":"education","lineage":["https://openalex.org/I170215575"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Kele Xu","raw_affiliation_strings":["National University of Defense Technology, Changsha, China and State Key Laboratory of Complex & Critical Software Environment, Changsha, China"],"affiliations":[{"raw_affiliation_string":"National University of Defense Technology, Changsha, China and State Key Laboratory of Complex & Critical Software Environment, Changsha, China","institution_ids":["https://openalex.org/I170215575"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100397991","display_name":"Yue Yu","orcid":"https://orcid.org/0000-0002-9865-2212"},"institutions":[{"id":"https://openalex.org/I170215575","display_name":"National University of Defense Technology","ror":"https://ror.org/05d2yfz11","country_code":"CN","type":"education","lineage":["https://openalex.org/I170215575"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Yue Yu","raw_affiliation_strings":["National University of Defense Technology, Changsha, China and State Key Laboratory of Complex & Critical Software Environment, Changsha, China"],"affiliations":[{"raw_affiliation_string":"National University of Defense Technology, Changsha, China and State Key Laboratory of Complex & Critical Software Environment, Changsha, China","institution_ids":["https://openalex.org/I170215575"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5108911205","display_name":"Wei Li","orcid":"https://orcid.org/0009-0007-0450-6075"},"institutions":[{"id":"https://openalex.org/I198357462","display_name":"Changsha University","ror":"https://ror.org/011d8sm39","country_code":"CN","type":"education","lineage":["https://openalex.org/I198357462"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Wei Li","raw_affiliation_strings":["Independent Researcher, Changsha, China"],"affiliations":[{"raw_affiliation_string":"Independent Researcher, Changsha, China","institution_ids":["https://openalex.org/I198357462"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100731006","display_name":"Hui Wang","orcid":"https://orcid.org/0000-0002-0499-727X"},"institutions":[{"id":"https://openalex.org/I170215575","display_name":"National University of Defense Technology","ror":"https://ror.org/05d2yfz11","country_code":"CN","type":"education","lineage":["https://openalex.org/I170215575"]},{"id":"https://openalex.org/I4210136793","display_name":"Peng Cheng Laboratory","ror":"https://ror.org/03qdqbt06","country_code":"CN","type":"facility","lineage":["https://openalex.org/I4210136793"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Hui Wang","raw_affiliation_strings":["National University of Defense Technology, Changsha, China and State Key Laboratory of Complex & Critical Software Environment, Changsha, China","Peng Cheng Laboratory, Shenzhen, China"],"affiliations":[{"raw_affiliation_string":"National University of Defense Technology, Changsha, China and State Key Laboratory of Complex & Critical Software Environment, Changsha, China","institution_ids":["https://openalex.org/I170215575"]},{"raw_affiliation_string":"Peng Cheng Laboratory, Shenzhen, China","institution_ids":["https://openalex.org/I4210136793"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5051227924","display_name":"Zenglin Xu","orcid":"https://orcid.org/0000-0001-5550-6461"},"institutions":[{"id":"https://openalex.org/I4210136793","display_name":"Peng Cheng Laboratory","ror":"https://ror.org/03qdqbt06","country_code":"CN","type":"facility","lineage":["https://openalex.org/I4210136793"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zenglin Xu","raw_affiliation_strings":["Peng Cheng Laboratory, Shenzhen, China and Fudan University, Shanghai, China"],"affiliations":[{"raw_affiliation_string":"Peng Cheng Laboratory, Shenzhen, China and Fudan University, Shanghai, China","institution_ids":["https://openalex.org/I4210136793"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5039795290","display_name":"Dawei Feng","orcid":"https://orcid.org/0000-0002-7587-8905"},"institutions":[{"id":"https://openalex.org/I170215575","display_name":"National University of Defense Technology","ror":"https://ror.org/05d2yfz11","country_code":"CN","type":"education","lineage":["https://openalex.org/I170215575"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Dawei Feng","raw_affiliation_strings":["National University of Defense Technology, Changsha, China and State Key Laboratory of Complex & Critical Software Environment, Changsha, China"],"affiliations":[{"raw_affiliation_string":"National University of Defense Technology, Changsha, China and State Key Laboratory of Complex & Critical Software Environment, Changsha, China","institution_ids":["https://openalex.org/I170215575"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5088885490","display_name":"Bo Ding","orcid":"https://orcid.org/0000-0002-1236-8318"},"institutions":[{"id":"https://openalex.org/I170215575","display_name":"National University of Defense Technology","ror":"https://ror.org/05d2yfz11","country_code":"CN","type":"education","lineage":["https://openalex.org/I170215575"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Bo Ding","raw_affiliation_strings":["National University of Defense Technology, Changsha, China and State Key Laboratory of Complex & Critical Software Environment, Changsha, China"],"affiliations":[{"raw_affiliation_string":"National University of Defense Technology, Changsha, China and State Key Laboratory of Complex & Critical Software Environment, Changsha, China","institution_ids":["https://openalex.org/I170215575"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5101522100","display_name":"Huaimin Wang","orcid":"https://orcid.org/0000-0002-3245-1901"},"institutions":[{"id":"https://openalex.org/I170215575","display_name":"National University of Defense Technology","ror":"https://ror.org/05d2yfz11","country_code":"CN","type":"education","lineage":["https://openalex.org/I170215575"]},{"id":"https://openalex.org/I4210136793","display_name":"Peng Cheng Laboratory","ror":"https://ror.org/03qdqbt06","country_code":"CN","type":"facility","lineage":["https://openalex.org/I4210136793"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Huaimin Wang","raw_affiliation_strings":["National University of Defense Technology, Changsha, China and State Key Laboratory of Complex & Critical Software Environment, Changsha, China","Peng Cheng Laboratory, Shenzhen, China"],"affiliations":[{"raw_affiliation_string":"National University of Defense Technology, Changsha, China and State Key Laboratory of Complex & Critical Software Environment, Changsha, China","institution_ids":["https://openalex.org/I170215575"]},{"raw_affiliation_string":"Peng Cheng Laboratory, Shenzhen, China","institution_ids":["https://openalex.org/I4210136793"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":11,"corresponding_author_ids":["https://openalex.org/A5073132517"],"corresponding_institution_ids":["https://openalex.org/I170215575"],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.19974638,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"191","last_page":"200"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T12101","display_name":"Advanced Bandit Algorithms Research","score":0.9984999895095825,"subfield":{"id":"https://openalex.org/subfields/1803","display_name":"Management Science and Operations Research"},"field":{"id":"https://openalex.org/fields/18","display_name":"Decision Sciences"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}},"topics":[{"id":"https://openalex.org/T12101","display_name":"Advanced Bandit Algorithms Research","score":0.9984999895095825,"subfield":{"id":"https://openalex.org/subfields/1803","display_name":"Management Science and Operations Research"},"field":{"id":"https://openalex.org/fields/18","display_name":"Decision Sciences"},"domain":{"id":"https://openalex.org/domains/2","display_name":"Social Sciences"}},{"id":"https://openalex.org/T10462","display_name":"Reinforcement Learning in Robotics","score":0.9975000023841858,"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":0.9945999979972839,"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/preference","display_name":"Preference","score":0.7767933011054993},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.624445378780365},{"id":"https://openalex.org/keywords/generative-grammar","display_name":"Generative grammar","score":0.5902377367019653},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.4198530614376068},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.12931722402572632},{"id":"https://openalex.org/keywords/statistics","display_name":"Statistics","score":0.09072068333625793}],"concepts":[{"id":"https://openalex.org/C2781249084","wikidata":"https://www.wikidata.org/wiki/Q908656","display_name":"Preference","level":2,"score":0.7767933011054993},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.624445378780365},{"id":"https://openalex.org/C39890363","wikidata":"https://www.wikidata.org/wiki/Q36108","display_name":"Generative grammar","level":2,"score":0.5902377367019653},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4198530614376068},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.12931722402572632},{"id":"https://openalex.org/C105795698","wikidata":"https://www.wikidata.org/wiki/Q12483","display_name":"Statistics","level":1,"score":0.09072068333625793}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3726302.3730063","is_oa":false,"landing_page_url":"https://doi.org/10.1145/3726302.3730063","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 48th International ACM SIGIR Conference on Research and Development in Information Retrieval","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":8,"referenced_works":["https://openalex.org/W2990138404","https://openalex.org/W3180880273","https://openalex.org/W4226278401","https://openalex.org/W4389523706","https://openalex.org/W4400524512","https://openalex.org/W4400524801","https://openalex.org/W4400525725","https://openalex.org/W4404261122"],"related_works":["https://openalex.org/W4391375266","https://openalex.org/W2899084033","https://openalex.org/W2748952813","https://openalex.org/W2380075625","https://openalex.org/W2390279801","https://openalex.org/W4391913857","https://openalex.org/W2358668433","https://openalex.org/W4396701345","https://openalex.org/W2376932109","https://openalex.org/W2001405890"],"abstract_inverted_index":{"Self-improving":[0],"alignment":[1,192,218],"leveraging":[2],"large":[3],"language":[4],"models":[5],"(LLMs)":[6],"to":[7,42,91,102,124,146,160],"automatically":[8],"generate":[9],"synthetic":[10],"preference":[11,74,110,122,148,188],"data":[12],"has":[13],"garnered":[14],"significant":[15],"attention":[16],"as":[17,113,147],"a":[18,96,114,156,211],"means":[19],"of":[20,85,165],"reducing":[21,196],"reliance":[22],"on":[23,151,169,175,221],"human":[24,194],"labelers.":[25],"These":[26],"methods":[27,60],"typically":[28],"employ":[29],"the":[30,34,49,54,83,109,120,133,141,163,172],"LLM-as-a-judge":[31,66],"mechanism,":[32],"where":[33],"LLM":[35,217],"generates":[36],"responses":[37],"and":[38,76,136,199,207,213],"then":[39],"employs":[40],"itself":[41],"judge":[43],"which":[44],"response":[45,135,167,177],"best":[46],"aligns":[47],"with":[48,193],"given":[50],"prompt":[51],"for":[52,216],"curating":[53],"binary":[55,88],"self-preferred":[56],"dataset.":[57],"However,":[58],"these":[59,104],"encounter":[61],"two":[62],"major":[63],"challenges:":[64],"(1)":[65],"often":[67,81],"produces":[68],"error-prone":[69],"evaluations,":[70],"resulting":[71],"in":[72,187],"low-quality":[73],"annotation,":[75],"(2)":[77],"their":[78],"optimization":[79],"strategies":[80],"overlook":[82],"strength":[84],"preferences":[86],"within":[87],"pairs,":[89],"leading":[90],"overfitting.":[92,201],"This":[93],"paper":[94],"proposes":[95],"novel":[97],"method,":[98],"Preference-Strength-aware":[99],"Optimization":[100],"(PSO),":[101],"address":[103],"issues.":[105],"Specifically,":[106],"PSO":[107,183,203],"frames":[108],"annotation":[111],"process":[112],"judgment":[115,130],"token":[116,131],"prediction":[117],"task":[118],"using":[119],"generative":[121],"model":[123],"produce":[125],"reliable":[126],"judgments.":[127],"The":[128],"predicted":[129],"indicates":[132],"preferred":[134],"its":[137],"corresponding":[138],"probability":[139],"reflects":[140],"disparity":[142],"between":[143],"responses,":[144,198],"referred":[145],"strength.":[149],"Based":[150],"this":[152],"strength,":[153],"we":[154],"introduce":[155],"new":[157],"preference-strength-aware":[158],"loss":[159],"adaptively":[161],"reweight":[162],"impact":[164],"different":[166],"pairs":[168],"optimization,":[170],"concentrating":[171],"model's":[173],"learning":[174],"high-quality":[176],"pairs.":[178],"Our":[179],"experiments":[180],"demonstrate":[181],"that":[182],"significantly":[184],"improves":[185],"performance":[186],"benchmarks,":[189],"achieving":[190],"stronger":[191],"preferences,":[195],"verbose":[197],"mitigating":[200],"Furthermore,":[202],"exhibits":[204],"robust":[205],"generalization":[206],"sample":[208],"efficiency,":[209],"offering":[210],"scalable":[212],"promising":[214],"solution":[215],"without":[219],"relying":[220],"human-annotated":[222],"preferences.":[223]},"counts_by_year":[],"updated_date":"2026-03-27T05:58:40.876381","created_date":"2025-10-10T00:00:00"}