{"id":"https://openalex.org/W4408355378","doi":"https://doi.org/10.1109/icassp49660.2025.10890215","title":"Found In The Distribution: Utilizing Latent Dirichlet Allocation Improves Long Context Comprehension of Large Language Models","display_name":"Found In The Distribution: Utilizing Latent Dirichlet Allocation Improves Long Context Comprehension of Large Language Models","publication_year":2025,"publication_date":"2025-03-12","ids":{"openalex":"https://openalex.org/W4408355378","doi":"https://doi.org/10.1109/icassp49660.2025.10890215"},"language":"en","primary_location":{"id":"doi:10.1109/icassp49660.2025.10890215","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp49660.2025.10890215","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ICASSP 2025 - 2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)","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/A5033846632","display_name":"Zhenyu Guan","orcid":"https://orcid.org/0009-0005-7601-7380"},"institutions":[{"id":"https://openalex.org/I78988378","display_name":"Renmin University of China","ror":"https://ror.org/041pakw92","country_code":"CN","type":"education","lineage":["https://openalex.org/I78988378"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Zhenyu Guan","raw_affiliation_strings":["Renmin University of China,CHN,School of Information,Beijing,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Renmin University of China,CHN,School of Information,Beijing,China","institution_ids":["https://openalex.org/I78988378"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5020649203","display_name":"Xun Liang","orcid":"https://orcid.org/0000-0002-3431-5954"},"institutions":[{"id":"https://openalex.org/I78988378","display_name":"Renmin University of China","ror":"https://ror.org/041pakw92","country_code":"CN","type":"education","lineage":["https://openalex.org/I78988378"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Xun Liang","raw_affiliation_strings":["Renmin University of China,CHN,School of Information,Beijing,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Renmin University of China,CHN,School of Information,Beijing,China","institution_ids":["https://openalex.org/I78988378"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5101600324","display_name":"Sensen Zhang","orcid":"https://orcid.org/0000-0002-0449-4699"},"institutions":[{"id":"https://openalex.org/I78988378","display_name":"Renmin University of China","ror":"https://ror.org/041pakw92","country_code":"CN","type":"education","lineage":["https://openalex.org/I78988378"]}],"countries":["CN"],"is_corresponding":false,"raw_author_name":"Sensen Zhang","raw_affiliation_strings":["Renmin University of China,CHN,School of Information,Beijing,China"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Renmin University of China,CHN,School of Information,Beijing,China","institution_ids":["https://openalex.org/I78988378"]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":1,"corresponding_author_ids":[],"corresponding_institution_ids":["https://openalex.org/I78988378"],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.02756936,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"1","last_page":"5"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10028","display_name":"Topic Modeling","score":0.98580002784729,"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":0.98580002784729,"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":0.9724000096321106,"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/latent-dirichlet-allocation","display_name":"Latent Dirichlet allocation","score":0.9151039123535156},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.6944578886032104},{"id":"https://openalex.org/keywords/context","display_name":"Context (archaeology)","score":0.5601445436477661},{"id":"https://openalex.org/keywords/comprehension","display_name":"Comprehension","score":0.5281928777694702},{"id":"https://openalex.org/keywords/hierarchical-dirichlet-process","display_name":"Hierarchical Dirichlet process","score":0.4936903417110443},{"id":"https://openalex.org/keywords/natural-language-processing","display_name":"Natural language processing","score":0.46122199296951294},{"id":"https://openalex.org/keywords/distribution","display_name":"Distribution (mathematics)","score":0.42281436920166016},{"id":"https://openalex.org/keywords/topic-model","display_name":"Topic model","score":0.4163423776626587},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.38612014055252075},{"id":"https://openalex.org/keywords/mathematics","display_name":"Mathematics","score":0.1603480875492096},{"id":"https://openalex.org/keywords/history","display_name":"History","score":0.10716769099235535},{"id":"https://openalex.org/keywords/programming-language","display_name":"Programming language","score":0.08708503842353821}],"concepts":[{"id":"https://openalex.org/C500882744","wikidata":"https://www.wikidata.org/wiki/Q269236","display_name":"Latent Dirichlet allocation","level":3,"score":0.9151039123535156},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6944578886032104},{"id":"https://openalex.org/C2779343474","wikidata":"https://www.wikidata.org/wiki/Q3109175","display_name":"Context (archaeology)","level":2,"score":0.5601445436477661},{"id":"https://openalex.org/C511192102","wikidata":"https://www.wikidata.org/wiki/Q5156948","display_name":"Comprehension","level":2,"score":0.5281928777694702},{"id":"https://openalex.org/C141318989","wikidata":"https://www.wikidata.org/wiki/Q5753066","display_name":"Hierarchical Dirichlet process","level":4,"score":0.4936903417110443},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.46122199296951294},{"id":"https://openalex.org/C110121322","wikidata":"https://www.wikidata.org/wiki/Q865811","display_name":"Distribution (mathematics)","level":2,"score":0.42281436920166016},{"id":"https://openalex.org/C171686336","wikidata":"https://www.wikidata.org/wiki/Q3532085","display_name":"Topic model","level":2,"score":0.4163423776626587},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.38612014055252075},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.1603480875492096},{"id":"https://openalex.org/C95457728","wikidata":"https://www.wikidata.org/wiki/Q309","display_name":"History","level":0,"score":0.10716769099235535},{"id":"https://openalex.org/C199360897","wikidata":"https://www.wikidata.org/wiki/Q9143","display_name":"Programming language","level":1,"score":0.08708503842353821},{"id":"https://openalex.org/C166957645","wikidata":"https://www.wikidata.org/wiki/Q23498","display_name":"Archaeology","level":1,"score":0.0},{"id":"https://openalex.org/C134306372","wikidata":"https://www.wikidata.org/wiki/Q7754","display_name":"Mathematical analysis","level":1,"score":0.0}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1109/icassp49660.2025.10890215","is_oa":false,"landing_page_url":"https://doi.org/10.1109/icassp49660.2025.10890215","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"ICASSP 2025 - 2025 IEEE International Conference on Acoustics, Speech and Signal Processing (ICASSP)","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":22,"referenced_works":["https://openalex.org/W2101105183","https://openalex.org/W2896457183","https://openalex.org/W2951434086","https://openalex.org/W2963926728","https://openalex.org/W3099700870","https://openalex.org/W3156789018","https://openalex.org/W4253028564","https://openalex.org/W4391876619","https://openalex.org/W4393145623","https://openalex.org/W4402684028","https://openalex.org/W6682631176","https://openalex.org/W6766673545","https://openalex.org/W6769627184","https://openalex.org/W6771915120","https://openalex.org/W6781533629","https://openalex.org/W6793886675","https://openalex.org/W6811129797","https://openalex.org/W6854962165","https://openalex.org/W6855665169","https://openalex.org/W6855750478","https://openalex.org/W6858095362","https://openalex.org/W6860465438"],"related_works":["https://openalex.org/W2097627380","https://openalex.org/W2914864478","https://openalex.org/W2008338582","https://openalex.org/W4291700620","https://openalex.org/W1999586157","https://openalex.org/W22044811","https://openalex.org/W2625329765","https://openalex.org/W2810994508","https://openalex.org/W2255612897","https://openalex.org/W2955328590"],"abstract_inverted_index":{"Large":[0],"Language":[1],"Models,":[2],"even":[3,99],"when":[4],"specifically":[5],"trained":[6],"to":[7,13,50,62,84,90],"process":[8],"long":[9,120],"input":[10],"contexts,":[11],"struggle":[12],"capture":[14,63],"relevant":[15],"information":[16,66,118],"located":[17],"in":[18,104,115],"the":[19,29,52,60,64,68,72,86,105],"middle":[20],"of":[21,54,97],"their":[22],"input.":[23],"This":[24],"phenomenon":[25],"is":[26],"known":[27],"as":[28],"\"lost-in-the-middle\"":[30],"problem.":[31],"In":[32,41],"this":[33],"study,":[34],"We":[35],"propose":[36],"a":[37,81],"new":[38],"method":[39,83],"Found":[40],"The":[42],"Distribution":[43],"(FITD)":[44],"with":[45],"Latent":[46],"Dirichlet":[47],"Allocation":[48],"(LDA)":[49],"solve":[51],"problem":[53],"\"lost-in-the-middle\".":[55],"Specifically,":[56],"we":[57,75,108],"first":[58],"apply":[59],"LDA":[61],"critical":[65],"within":[67,119],"text.":[69],"Based":[70],"on":[71,93],"captured":[73],"results,":[74],"generate":[76],"corresponding":[77],"probabilities":[78],"and":[79],"use":[80],"masking":[82],"compress":[85],"prompt,":[87],"enabling":[88],"LLMs":[89],"correctly":[91],"focus":[92],"these":[94],"important":[95],"pieces":[96],"information,":[98],"if":[100],"they":[101],"are":[102],"situated":[103],"middle.":[106],"Finally,":[107],"demonstrate":[109],"that":[110],"our":[111],"approach":[112],"performs":[113],"better":[114],"identifying":[116],"pertinent":[117],"contexts.":[121]},"counts_by_year":[],"updated_date":"2026-07-29T14:22:42.915294","created_date":"2025-10-10T00:00:00"}