{"id":"https://openalex.org/W7166890842","doi":"https://doi.org/10.18653/v1/2026.findings-acl.1520","title":"Exploration-Exploitation Reshaping towards Efficient Reasoning for Large Language Models","display_name":"Exploration-Exploitation Reshaping towards Efficient Reasoning for Large Language Models","publication_year":2026,"publication_date":"2026-01-01","ids":{"openalex":"https://openalex.org/W7166890842","doi":"https://doi.org/10.18653/v1/2026.findings-acl.1520"},"language":null,"primary_location":{"id":"doi:10.18653/v1/2026.findings-acl.1520","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-acl.1520","pdf_url":"https://aclanthology.org/2026.findings-acl.1520.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":"Findings of the Association for Computational Linguistics: ACL 2026","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://aclanthology.org/2026.findings-acl.1520.pdf","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5139721472","display_name":"Yufeng Shi","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yufeng Shi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139739068","display_name":"Weilin Luo","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Weilin Luo","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139756448","display_name":"Yuxiang Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yuxiang Zhang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139781688","display_name":"Zongmeng Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zongmeng Zhang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139740893","display_name":"Haoyang Liu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Haoyang Liu","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139724629","display_name":"Yubing Wang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yubing Wang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139724137","display_name":"Wang Bin","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Bin Wang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5139767097","display_name":"Wengang Zhou","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wengang Zhou","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5139807237","display_name":"Houqiang Li","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Houqiang Li","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":0.8787903,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"30392","last_page":"30407"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10181","display_name":"Natural Language Processing Techniques","score":0.18469999730587006,"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/T10181","display_name":"Natural Language Processing Techniques","score":0.18469999730587006,"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":0.17630000412464142,"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.1168999969959259,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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/automated-reasoning","display_name":"Automated reasoning","score":0.353300005197525},{"id":"https://openalex.org/keywords/natural-language","display_name":"Natural language","score":0.3384000062942505},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.3190999925136566},{"id":"https://openalex.org/keywords/semantics","display_name":"Semantics (computer science)","score":0.2856000065803528},{"id":"https://openalex.org/keywords/non-monotonic-logic","display_name":"Non-monotonic logic","score":0.2590000033378601}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6525999903678894},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.48330000042915344},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.4207000136375427},{"id":"https://openalex.org/C195344581","wikidata":"https://www.wikidata.org/wiki/Q2555318","display_name":"Automated reasoning","level":2,"score":0.353300005197525},{"id":"https://openalex.org/C195324797","wikidata":"https://www.wikidata.org/wiki/Q33742","display_name":"Natural language","level":2,"score":0.3384000062942505},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.3190999925136566},{"id":"https://openalex.org/C184337299","wikidata":"https://www.wikidata.org/wiki/Q1437428","display_name":"Semantics (computer science)","level":2,"score":0.2856000065803528},{"id":"https://openalex.org/C41895202","wikidata":"https://www.wikidata.org/wiki/Q8162","display_name":"Linguistics","level":1,"score":0.2825999855995178},{"id":"https://openalex.org/C159032336","wikidata":"https://www.wikidata.org/wiki/Q2488768","display_name":"Non-monotonic logic","level":2,"score":0.2590000033378601},{"id":"https://openalex.org/C2776401178","wikidata":"https://www.wikidata.org/wiki/Q12050496","display_name":"Feature (linguistics)","level":2,"score":0.25780001282691956},{"id":"https://openalex.org/C161301231","wikidata":"https://www.wikidata.org/wiki/Q3478658","display_name":"Knowledge representation and reasoning","level":2,"score":0.25679999589920044}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.18653/v1/2026.findings-acl.1520","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-acl.1520","pdf_url":"https://aclanthology.org/2026.findings-acl.1520.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":"Findings of the Association for Computational Linguistics: ACL 2026","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.18653/v1/2026.findings-acl.1520","is_oa":true,"landing_page_url":"https://doi.org/10.18653/v1/2026.findings-acl.1520","pdf_url":"https://aclanthology.org/2026.findings-acl.1520.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":"Findings of the Association for Computational Linguistics: ACL 2026","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7166890842.pdf","grobid_xml":"https://content.openalex.org/works/W7166890842.grobid-xml"},"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"While":[0],"excelling":[1],"at":[2],"solving":[3],"complex":[4],"problems,":[5],"Large":[6],"Reasoning":[7],"Models":[8],"(LRMs)":[9],"are":[10],"still":[11],"constrained":[12],"by":[13,87],"the":[14,29,40,66,79,132,139,148,151],"overthinking":[15],"issue.Most":[16],"current":[17],"studies":[18,136],"rely":[19],"on":[20,98],"reward":[21],"shaping":[22],"in":[23],"Reinforcement":[24],"Learning":[25],"(RL)":[26],"to":[27,39],"shorten":[28],"Chainof-Thought":[30],"(CoT)":[31],"of":[32,42,82,156],"LRMs,":[33],"remaining":[34],"sampleinefficient":[35],"and":[36,45,60,68,85,92,112,129,154],"non-robust":[37],"due":[38],"absence":[41],"prioritized":[43,80],"exploitation":[44,69,81],"guided":[46],"exploration.To":[47],"address":[48],"these":[49],"issues,":[50],"we":[51],"propose":[52],"a":[53,89,93,106],"novel":[54],"policy":[55],"optimization":[56],"framework":[57],"with":[58],"Self-Imitation":[59],"self-Guidance":[61],"MechAnisms":[62],"(SIGMA),":[63],"which":[64,77,104],"reshapes":[65],"exploration":[67,108],"through":[70,110],"two":[71],"core":[72],"components:":[73],"(i)":[74],"self-imitation":[75,90],"exploitation,":[76],"enables":[78],"high-value":[83],"prompts":[84],"rollouts":[86],"introducing":[88],"loss":[91],"dynamic":[94],"sampling":[95],"strategy":[96],"based":[97],"compression":[99],"rate;":[100],"(ii)":[101],"self-guidance":[102],"exploration,":[103],"provides":[105],"preference-aware":[107],"guidance":[109],"diverse":[111],"pluggable":[113],"selfrewriting":[114],"strategies.Experiments":[115],"across":[116],"various":[117],"datasets":[118],"indicate":[119],"that":[120,138],"our":[121],"method":[122],"achieves":[123],"superior":[124],"reasoning":[125,152],"efficiency":[126,155],"without":[127],"compromising,":[128],"even":[130],"facilitating,":[131],"overall":[133],"accuracy.Furthermore,":[134],"ablation":[135],"show":[137],"proposed":[140],"mechanisms":[141],"can":[142],"provide":[143],"flexible":[144],"control":[145],"interfaces":[146],"for":[147],"tradeoff":[149],"between":[150],"accuracy":[153],"LRMs.*Equal":[157],"contribution.":[158]},"counts_by_year":[],"updated_date":"2026-07-29T14:22:42.915294","created_date":"2026-07-02T00:00:00"}