{"id":"https://openalex.org/W7207552497","doi":"https://doi.org/10.48550/arxiv.2609.01913","title":"MERGED: Multimodal Entity Resolution via Generated Expert Reasoning Distillation","display_name":"MERGED: Multimodal Entity Resolution via Generated Expert Reasoning Distillation","publication_year":2026,"publication_date":"2026-09-01","ids":{"openalex":"https://openalex.org/W7207552497","doi":"https://doi.org/10.48550/arxiv.2609.01913"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2609.01913","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.01913","pdf_url":"https://arxiv.org/pdf/2609.01913","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"volume":null,"type":"preprint","is_traditional":true,"indexed_in":["arxiv","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2609.01913","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5127706979","display_name":"You-Lin Chen","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chen, You-Lin","raw_affiliation_strings":[", United States"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":", United States","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5149750036","display_name":"Kyoungjun Park","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Park, Kyoungjun","raw_affiliation_strings":[", United States"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":", United States","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5121541199","display_name":"Bin Xu","orcid":"https://orcid.org/0000-0001-5732-7985","observed_orcids":["https://orcid.org/0000-0001-5732-7985"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xu, Bin","raw_affiliation_strings":[", United States"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":", United States","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5105422045","display_name":"Prithviraj Sen","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Sen, Prithviraj","raw_affiliation_strings":[", United States"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":", United States","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5158144345","display_name":"Abhishek Prasad Ram Tripathi","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Tripathi, Abhishek Prasad Ram","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5114453044","display_name":"Pedro Herrero-Vidal","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Herrero-Vidal, Pedro","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":true,"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/T11719","display_name":"Data Quality and Management","score":0.8095999956130981,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"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/T11719","display_name":"Data Quality and Management","score":0.8095999956130981,"subfield":{"id":"https://openalex.org/subfields/1710","display_name":"Information Systems"},"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":0.12880000472068787,"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":0.02879999950528145,"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/knowledge-distillation","display_name":"knowledge distillation","score":0.8700000047683716},{"id":"https://openalex.org/keywords/vision-language-models","display_name":"vision-language models","score":0.9070000052452087},{"id":"https://openalex.org/keywords/reasoning-distillation","display_name":"reasoning distillation","score":0.7630000114440918},{"id":"https://openalex.org/keywords/direct-preference-optimization","display_name":"direct preference optimization","score":0.6970000267028809},{"id":"https://openalex.org/keywords/zero-shot-transfer","display_name":"zero-shot transfer","score":0.4090000092983246}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7382000088691711},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.6158000230789185},{"id":"https://openalex.org/C82876162","wikidata":"https://www.wikidata.org/wiki/Q17096504","display_name":"Latency (audio)","level":2,"score":0.4896000027656555},{"id":"https://openalex.org/C146849305","wikidata":"https://www.wikidata.org/wiki/Q370766","display_name":"Ground truth","level":2,"score":0.43650001287460327},{"id":"https://openalex.org/C139807058","wikidata":"https://www.wikidata.org/wiki/Q352374","display_name":"Adaptation (eye)","level":2,"score":0.43549999594688416},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.4268999993801117},{"id":"https://openalex.org/C2776321320","wikidata":"https://www.wikidata.org/wiki/Q857525","display_name":"Annotation","level":2,"score":0.42559999227523804},{"id":"https://openalex.org/C138268822","wikidata":"https://www.wikidata.org/wiki/Q1051925","display_name":"Resolution (logic)","level":2,"score":0.4088999927043915},{"id":"https://openalex.org/C90673727","wikidata":"https://www.wikidata.org/wiki/Q901718","display_name":"Product (mathematics)","level":2,"score":0.4083999991416931},{"id":"https://openalex.org/C12725497","wikidata":"https://www.wikidata.org/wiki/Q810247","display_name":"Baseline (sea)","level":2,"score":0.39239999651908875},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.33149999380111694},{"id":"https://openalex.org/C195344581","wikidata":"https://www.wikidata.org/wiki/Q2555318","display_name":"Automated reasoning","level":2,"score":0.32350000739097595},{"id":"https://openalex.org/C2777877512","wikidata":"https://www.wikidata.org/wiki/Q1116097","display_name":"Common ground","level":2,"score":0.311599999666214},{"id":"https://openalex.org/C2781249084","wikidata":"https://www.wikidata.org/wiki/Q908656","display_name":"Preference","level":2,"score":0.31139999628067017},{"id":"https://openalex.org/C46637626","wikidata":"https://www.wikidata.org/wiki/Q6693015","display_name":"Low latency (capital markets)","level":2,"score":0.25920000672340393},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.25369998812675476}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2609.01913","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.01913","pdf_url":"https://arxiv.org/pdf/2609.01913","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2609.01913","doi":"https://doi.org/10.48550/arxiv.2609.01913","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2609.01913","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2609.01913","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2609.01913","pdf_url":"https://arxiv.org/pdf/2609.01913","source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":"2331-8422","issn":["2331-8422"],"is_oa":true,"is_in_doaj":false,"is_core":false,"listed_in":[],"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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7207552497.pdf","grobid_xml":"https://content.openalex.org/works/W7207552497.grobid-xml"},"referenced_works_count":6,"referenced_works":["https://openalex.org/W4403578740","https://openalex.org/W4406491470","https://openalex.org/W4406961529","https://openalex.org/W4407231762","https://openalex.org/W4414900373","https://openalex.org/W4417287136"],"related_works":["https://openalex.org/W7128061673","https://openalex.org/W7166655313","https://openalex.org/W4385571924","https://openalex.org/W4416016938","https://openalex.org/W4416037389","https://openalex.org/W7160044430","https://openalex.org/W4401639961","https://openalex.org/W3173169192","https://openalex.org/W4403585694","https://openalex.org/W4392127960"],"abstract_inverted_index":{"In":[0],"product":[1,94],"entity":[2],"resolution,":[3],"relationship":[4,180,202],"definitions":[5],"constantly":[6],"evolve":[7],"with":[8,182],"business":[9],"needs,":[10],"yet":[11],"adapting":[12],"to":[13,38,177,200],"each":[14,93],"change":[15],"traditionally":[16],"requires":[17],"slow,":[18],"costly":[19],"human":[20,48,86,145],"annotation":[21,87],"that":[22,47,67],"is":[23],"often":[24],"noisy":[25],"and":[26,43,54,96,147,192,216],"carries":[27],"no":[28,85],"reasoning.":[29],"Large":[30],"vision-language":[31],"models":[32],"(VLMs)":[33],"prompted":[34],"zero-shot":[35,191],"can":[36],"adapt":[37],"a":[39,64,80,113,128,178,205,214],"new":[40,179,206],"definition":[41,181],"immediately":[42],"supply":[44,105],"the":[45,98,132,140,149],"reasoning":[46,74,99],"labels":[49,71,146],"lack,":[50],"but":[51,72],"their":[52,101],"cost":[53,215],"latency":[55,217],"are":[56,110],"prohibitive":[57],"at":[58,155,213],"production":[59],"scale.":[60],"We":[61],"present":[62],"MERGED,":[63],"distillation":[65],"framework":[66],"transfers":[68],"not":[69],"just":[70],"structured":[73],"from":[75,172],"large":[76],"teacher":[77],"VLMs":[78],"into":[79,115],"compact":[81],"7B-parameter":[82],"student,":[83],"requiring":[84],"for":[88,118,219],"training.":[89,195],"Multiple":[90],"teachers":[91],"label":[92],"pair":[95],"articulate":[97],"behind":[100],"decision:":[102],"agreement":[103],"pairs":[104,117],"supervised":[106],"fine-tuning,":[107],"while":[108,159],"disagreements":[109],"resolved":[111],"by":[112,137,153,188],"meta-judge":[114],"preference":[116],"Direct":[119],"Preference":[120],"Optimization.":[121],"Evaluated":[122],"against":[123],"human-labeled":[124],"ground":[125],"truth":[126],"on":[127,144],"multilingual":[129],"e-commerce":[130],"dataset,":[131],"resulting":[133],"student":[134],"improves":[135],"PR-AUC":[136,187],"13.79%":[138],"over":[139,190],"same":[141],"backbone":[142],"trained":[143],"surpasses":[148],"larger":[150],"Qwen2.5-32B-VL":[151],"baseline":[152],"6.32%":[154],"6x":[156],"lower":[157],"cost,":[158],"also":[160],"yielding":[161],"tighter":[162],"label-reasoning":[163],"consistency":[164],"(over":[165],"10%":[166],"above":[167],"Qwen2.5-32B-VL).":[168],"Moreover,":[169],"re-applying":[170],"MERGED":[171,196],"an":[173],"existing":[174],"checkpoint":[175],"adapts":[176],"only":[183],"10K":[184],"samples,":[185],"improving":[186],"6.97%":[189],"outperforming":[193],"from-scratch":[194],"enables":[197],"rapid":[198],"adaptation":[199],"evolving":[201],"definitions,":[203],"supporting":[204],"one":[207],"in":[208],"days":[209],"rather":[210],"than":[211],"months,":[212],"suitable":[218],"large-scale":[220],"industrial":[221],"deployment.":[222]},"counts_by_year":[],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2026-09-04T00:00:00"}