{"id":"https://openalex.org/W7133488686","doi":"https://doi.org/10.48550/arxiv.2603.03010","title":"Reproducing and Comparing Distillation Techniques for Cross-Encoders","display_name":"Reproducing and Comparing Distillation Techniques for Cross-Encoders","publication_year":2026,"publication_date":"2026-03-03","ids":{"openalex":"https://openalex.org/W7133488686","doi":"https://doi.org/10.48550/arxiv.2603.03010"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2603.03010","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.03010","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"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":"article"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2603.03010","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5128067382","display_name":"Victor Morand","orcid":null},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Morand, Victor","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5093775107","display_name":"Mathias Vast","orcid":"https://orcid.org/0009-0007-4612-717X"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Vast, Mathias","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5093775108","display_name":"Basile Van Cooten","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Van Cooten, Basile","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5128102489","display_name":"Laure Soulier","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Soulier, Laure","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5104522309","display_name":"Josiane Mothe","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Mothe, Josiane","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5086752907","display_name":"Benjamin Piwowarski","orcid":"https://orcid.org/0000-0001-6792-3262"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Piwowarski, Benjamin","raw_affiliation_strings":[],"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":6,"corresponding_author_ids":["https://openalex.org/A5128067382"],"corresponding_institution_ids":[],"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":null,"last_page":null},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.07760000228881836,"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"}},"topics":[{"id":"https://openalex.org/T10036","display_name":"Advanced Neural Network Applications","score":0.07760000228881836,"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/T12026","display_name":"Explainable Artificial Intelligence (XAI)","score":0.07450000196695328,"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/T11448","display_name":"Face recognition and analysis","score":0.06210000067949295,"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/pairwise-comparison","display_name":"Pairwise comparison","score":0.7287999987602234},{"id":"https://openalex.org/keywords/pointwise","display_name":"Pointwise","score":0.6780999898910522},{"id":"https://openalex.org/keywords/distillation","display_name":"Distillation","score":0.6044999957084656},{"id":"https://openalex.org/keywords/range","display_name":"Range (aeronautics)","score":0.5475999712944031},{"id":"https://openalex.org/keywords/scaling","display_name":"Scaling","score":0.5055000185966492},{"id":"https://openalex.org/keywords/training-set","display_name":"Training set","score":0.385699987411499},{"id":"https://openalex.org/keywords/ranking","display_name":"Ranking (information retrieval)","score":0.35499998927116394}],"concepts":[{"id":"https://openalex.org/C184898388","wikidata":"https://www.wikidata.org/wiki/Q1435712","display_name":"Pairwise comparison","level":2,"score":0.7287999987602234},{"id":"https://openalex.org/C2777984123","wikidata":"https://www.wikidata.org/wiki/Q9248237","display_name":"Pointwise","level":2,"score":0.6780999898910522},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6601999998092651},{"id":"https://openalex.org/C204030448","wikidata":"https://www.wikidata.org/wiki/Q101017","display_name":"Distillation","level":2,"score":0.6044999957084656},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.5817000269889832},{"id":"https://openalex.org/C204323151","wikidata":"https://www.wikidata.org/wiki/Q905424","display_name":"Range (aeronautics)","level":2,"score":0.5475999712944031},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5151000022888184},{"id":"https://openalex.org/C99844830","wikidata":"https://www.wikidata.org/wiki/Q102441924","display_name":"Scaling","level":2,"score":0.5055000185966492},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.385699987411499},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.358599990606308},{"id":"https://openalex.org/C189430467","wikidata":"https://www.wikidata.org/wiki/Q7293293","display_name":"Ranking (information retrieval)","level":2,"score":0.35499998927116394},{"id":"https://openalex.org/C134121241","wikidata":"https://www.wikidata.org/wiki/Q899301","display_name":"Yield (engineering)","level":2,"score":0.325300008058548},{"id":"https://openalex.org/C2780428219","wikidata":"https://www.wikidata.org/wiki/Q16952335","display_name":"Cover (algebra)","level":2,"score":0.32089999318122864},{"id":"https://openalex.org/C185798385","wikidata":"https://www.wikidata.org/wiki/Q1161707","display_name":"Benchmark (surveying)","level":2,"score":0.3061999976634979},{"id":"https://openalex.org/C2778755073","wikidata":"https://www.wikidata.org/wiki/Q10858537","display_name":"Scale (ratio)","level":2,"score":0.2833999991416931},{"id":"https://openalex.org/C165064840","wikidata":"https://www.wikidata.org/wiki/Q1321061","display_name":"Matching (statistics)","level":2,"score":0.2806999981403351},{"id":"https://openalex.org/C127705205","wikidata":"https://www.wikidata.org/wiki/Q5748245","display_name":"Heuristics","level":2,"score":0.27900001406669617}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2603.03010","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.03010","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"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":null,"raw_source_name":null,"raw_type":"article"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2603.03010","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.03010","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"is_oa":true,"is_in_doaj":false,"is_core":false,"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":"article"},"sustainable_development_goals":[{"display_name":"Quality Education","id":"https://metadata.un.org/sdg/4","score":0.5142620205879211}],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Recent":[0,14],"advances":[1],"in":[2,12,70,80],"Information":[3],"Retrieval":[4],"have":[5,16,67],"established":[6],"transformer-based":[7],"cross-encoders":[8,29],"as":[9,113,115],"a":[10,58,121],"keystone":[11],"IR.":[13],"studies":[15,56,79],"focused":[17],"on":[18,107,146],"knowledge":[19],"distillation":[20,46,98],"and":[21,100,130,135,150,153,158,171,182,185],"showed":[22],"that,":[23],"with":[24,41],"the":[25,32,127,138,195],"right":[26],"strategy,":[27],"traditional":[28],"could":[30],"reach":[31],"level":[33],"of":[34,36,61,77,110,124],"effectiveness":[35],"LLM":[37],"re-rankers.":[38],"Yet,":[39],"comparisons":[40,167],"previous":[42],"training":[43],"strategies,":[44],"including":[45],"from":[47,126],"strong":[48],"cross-encoder":[49,111],"teachers,":[50,112],"remain":[51],"unclear.":[52],"In":[53,91],"addition,":[54],"few":[55],"cover":[57],"similar":[59],"range":[60,123],"backbone":[62,196],"encoders,":[63],"while":[64],"substantial":[65],"improvements":[66],"been":[68],"made":[69],"this":[71,92],"area":[72],"since":[73],"BERT.":[74],"This":[75],"lack":[76],"comprehensive":[78],"controlled":[81],"environments":[82],"makes":[83],"it":[84,102],"difficult":[85],"to":[86,103,119,137,193],"identify":[87],"robust":[88],"design":[89],"choices.":[90],"work,":[93],"we":[94],"reproduce":[95],"\\citet{schlattRankDistiLLMClosingEffectiveness2025}":[96],"LLM-based":[97],"strategy":[99],"compare":[101],"\\citet{hofstatterImprovingEfficientNeural2020}":[104],"approach":[105],"based":[106],"an":[108],"ensemble":[109],"well":[114],"other":[116],"supervised":[117],"objectives,":[118],"fine-tune":[120],"large":[122],"cross-encoders,":[125],"original":[128],"BERT":[129],"its":[131],"follow-ups":[132],"RoBERTa,":[133],"ELECTRA":[134],"DeBERTa-v3,":[136],"more":[139],"recent":[140],"ModernBERT.":[141],"We":[142],"evaluate":[143],"all":[144,180],"models":[145],"both":[147],"in-domain":[148],"(TREC-DL":[149],"MS~MARCO":[151],"dev)":[152],"out-of-domain":[154],"datasets":[155],"(BEIR,":[156],"LoTTE,":[157],"Robust04).":[159],"Our":[160],"results":[161],"show":[162],"that":[163,186],"objectives":[164],"emphasizing":[165],"relative":[166],"--":[168,174],"pairwise":[169],"MarginMSE":[170],"listwise":[172],"InfoNCE":[173],"consistently":[175],"outperform":[176],"pointwise":[177],"baselines":[178],"across":[179],"backbones":[181],"evaluation":[183],"settings,":[184],"objective":[187],"choice":[188],"can":[189],"yield":[190],"gains":[191],"comparable":[192],"scaling":[194],"architecture.":[197]},"counts_by_year":[],"updated_date":"2026-03-05T07:36:02.291473","created_date":"2026-03-05T00:00:00"}