{"id":"https://openalex.org/W7130530031","doi":"https://doi.org/10.1109/ictc66702.2025.11389069","title":"Policy-based Word Subset Selection for Explaining Black-Box Language Models","display_name":"Policy-based Word Subset Selection for Explaining Black-Box Language Models","publication_year":2025,"publication_date":"2025-10-14","ids":{"openalex":"https://openalex.org/W7130530031","doi":"https://doi.org/10.1109/ictc66702.2025.11389069"},"language":null,"primary_location":{"id":"doi:10.1109/ictc66702.2025.11389069","is_oa":false,"landing_page_url":"https://doi.org/10.1109/ictc66702.2025.11389069","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 16th International Conference on Information and Communication Technology Convergence (ICTC)","raw_type":"proceedings-article"},"type":"conference-paper","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/A5083426795","display_name":"Minyoung Hwang","orcid":null},"institutions":[{"id":"https://openalex.org/I197347611","display_name":"Korea University","ror":"https://ror.org/047dqcg40","country_code":"KR","type":"education","lineage":["https://openalex.org/I197347611"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Minyoung Hwang","raw_affiliation_strings":["Korea University,Department of Artificial Intelligence,Seoul,Korea"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Korea University,Department of Artificial Intelligence,Seoul,Korea","institution_ids":["https://openalex.org/I197347611"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5126371398","display_name":"Seokhyun Lee","orcid":null},"institutions":[{"id":"https://openalex.org/I197347611","display_name":"Korea University","ror":"https://ror.org/047dqcg40","country_code":"KR","type":"education","lineage":["https://openalex.org/I197347611"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Seokhyun Lee","raw_affiliation_strings":["Korea University,Department of Artificial Intelligence,Seoul,Korea"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Korea University,Department of Artificial Intelligence,Seoul,Korea","institution_ids":["https://openalex.org/I197347611"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5126369812","display_name":"Changhee Lee","orcid":null},"institutions":[{"id":"https://openalex.org/I197347611","display_name":"Korea University","ror":"https://ror.org/047dqcg40","country_code":"KR","type":"education","lineage":["https://openalex.org/I197347611"]}],"countries":["KR"],"is_corresponding":false,"raw_author_name":"Changhee Lee","raw_affiliation_strings":["Korea University,Department of Artificial Intelligence,Seoul,Korea"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Korea University,Department of Artificial Intelligence,Seoul,Korea","institution_ids":["https://openalex.org/I197347611"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I197347611"],"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":"219","last_page":"224"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":0.7634000182151794,"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/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":0.7634000182151794,"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":0.04360000044107437,"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"}},{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.031199999153614044,"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/interpretability","display_name":"Interpretability","score":0.892300009727478},{"id":"https://openalex.org/keywords/leverage","display_name":"Leverage (statistics)","score":0.7436000108718872},{"id":"https://openalex.org/keywords/salient","display_name":"Salient","score":0.7084000110626221},{"id":"https://openalex.org/keywords/language-model","display_name":"Language model","score":0.6452000141143799},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.6200000047683716},{"id":"https://openalex.org/keywords/selection","display_name":"Selection (genetic algorithm)","score":0.5972999930381775},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.4844000041484833},{"id":"https://openalex.org/keywords/word","display_name":"Word (group theory)","score":0.4196999967098236}],"concepts":[{"id":"https://openalex.org/C2781067378","wikidata":"https://www.wikidata.org/wiki/Q17027399","display_name":"Interpretability","level":2,"score":0.892300009727478},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7498999834060669},{"id":"https://openalex.org/C153083717","wikidata":"https://www.wikidata.org/wiki/Q6535263","display_name":"Leverage (statistics)","level":2,"score":0.7436000108718872},{"id":"https://openalex.org/C2780719617","wikidata":"https://www.wikidata.org/wiki/Q1030752","display_name":"Salient","level":2,"score":0.7084000110626221},{"id":"https://openalex.org/C137293760","wikidata":"https://www.wikidata.org/wiki/Q3621696","display_name":"Language model","level":2,"score":0.6452000141143799},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.6200000047683716},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.6097999811172485},{"id":"https://openalex.org/C81917197","wikidata":"https://www.wikidata.org/wiki/Q628760","display_name":"Selection (genetic algorithm)","level":2,"score":0.5972999930381775},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.4844000041484833},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.4767000079154968},{"id":"https://openalex.org/C90805587","wikidata":"https://www.wikidata.org/wiki/Q10944557","display_name":"Word (group theory)","level":2,"score":0.4196999967098236},{"id":"https://openalex.org/C93959086","wikidata":"https://www.wikidata.org/wiki/Q6888345","display_name":"Model selection","level":2,"score":0.4016000032424927},{"id":"https://openalex.org/C2777601683","wikidata":"https://www.wikidata.org/wiki/Q6499736","display_name":"Vocabulary","level":2,"score":0.4011000096797943},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.3391999900341034},{"id":"https://openalex.org/C2983448237","wikidata":"https://www.wikidata.org/wiki/Q1078276","display_name":"Language understanding","level":2,"score":0.30630001425743103},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.28780001401901245},{"id":"https://openalex.org/C104122410","wikidata":"https://www.wikidata.org/wiki/Q1416406","display_name":"Network model","level":2,"score":0.27950000762939453},{"id":"https://openalex.org/C2779439875","wikidata":"https://www.wikidata.org/wiki/Q1078276","display_name":"Natural language understanding","level":3,"score":0.2671999931335449},{"id":"https://openalex.org/C114289077","wikidata":"https://www.wikidata.org/wiki/Q3284399","display_name":"Statistical model","level":2,"score":0.26420000195503235},{"id":"https://openalex.org/C195324797","wikidata":"https://www.wikidata.org/wiki/Q33742","display_name":"Natural language","level":2,"score":0.2578999996185303}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/ictc66702.2025.11389069","is_oa":false,"landing_page_url":"https://doi.org/10.1109/ictc66702.2025.11389069","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 16th International Conference on Information and Communication Technology Convergence (ICTC)","raw_type":"proceedings-article"}],"best_oa_location":null,"sustainable_development_goals":[],"awards":[],"funders":[{"id":"https://openalex.org/F4320322120","display_name":"National Research Foundation of Korea","ror":"https://ror.org/013aysd81"}],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":7,"referenced_works":["https://openalex.org/W1849277567","https://openalex.org/W2091144449","https://openalex.org/W2336525064","https://openalex.org/W2516809705","https://openalex.org/W2782893163","https://openalex.org/W4241811150","https://openalex.org/W4391528766"],"related_works":[],"abstract_inverted_index":{"The":[0],"development":[1],"of":[2,128,146,162],"deep":[3],"language":[4],"models":[5,28,104],"(DLMs)":[6],"has":[7,14],"made":[8],"them":[9],"more":[10],"widely":[11],"used.":[12],"This":[13,41],"led":[15],"to":[16,34,87,139,186,201],"a":[17,46,60,115,123,135,141,159],"growing":[18],"need":[19],"for":[20,167,179],"tools":[21],"that":[22,117],"can":[23],"help":[24],"understand":[25],"how":[26],"these":[27,111],"work,":[29],"especially":[30,63],"when":[31,64],"it":[32,198],"comes":[33],"understanding":[35],"the":[36,165,193,208],"reasoning":[37],"behind":[38],"their":[39],"outputs.":[40],"explainability":[42],"is":[43],"emerging":[44],"as":[45,164],"key":[47],"factor":[48],"in":[49],"building":[50],"trust":[51],"between":[52],"users":[53],"and":[54,75,98,125],"technologies.":[55],"Achieving":[56],"meaningful":[57],"interpretability":[58],"remains":[59],"significant":[61],"challenge,":[62],"DLMs":[65],"are":[66,77],"considered":[67],"black-box":[68,103,202],"systems,":[69],"where":[70],"internal":[71],"details":[72],"like":[73],"parameters":[74],"gradients":[76,178],"inaccessible.":[78],"Although":[79],"many":[80],"techniques":[81],"have":[82],"been":[83],"proposed,":[84],"most":[85],"struggle":[86],"meet":[88],"two":[89],"critical":[90],"goals":[91],"simultaneously:":[92],"(i)":[93],"maintaining":[94],"efficiency":[95],"during":[96],"inference,":[97],"(ii)":[99],"remaining":[100],"compatible":[101],"with":[102,154,207],"without":[105,188,216],"causing":[106],"out-of-distribution":[107],"behaviors.":[108],"To":[109],"overcome":[110],"limitations,":[112],"we":[113,175],"introduce":[114],"method":[116,185],"explains":[118],"model":[119],"predictions":[120,215],"by":[121,204],"selecting":[122],"concise":[124],"informative":[126,144],"subset":[127,145,161],"input":[129,147],"words.":[130],"Our":[131],"approach":[132],"involves":[133],"training":[134],"lightweight":[136],"selection":[137],"network":[138,152],"identify":[140],"minimal":[142],"yet":[143],"tokens.":[148],"Once":[149],"trained,":[150],"this":[151],"operates":[153],"high":[155],"efficiency,":[156],"directly":[157,205],"identifying":[158],"salient":[160],"words":[163],"explanation":[166],"new":[168],"samples":[169],"at":[170],"inference":[171],"time.":[172],"For":[173],"training,":[174],"leverage":[176],"policy":[177],"optimization,":[180],"which":[181],"critically":[182],"allows":[183],"our":[184],"operate":[187],"requiring":[189,217],"gradient":[190,219],"information":[191],"from":[192],"target":[194,209],"DLM,":[195],"thus":[196],"making":[197],"inherently":[199],"applicable":[200],"systems":[203],"interacting":[206],"DLM":[210],"solely":[211],"through":[212],"its":[213],"input-output":[214],"any":[218],"information.":[220]},"counts_by_year":[],"updated_date":"2026-07-14T23:27:15.235271","created_date":"2026-02-20T00:00:00"}