{"id":"https://openalex.org/W7164560501","doi":"https://doi.org/10.48550/arxiv.2606.13061","title":"LaME: Learning to Think in Latent Space for Multimodal Embedding via Information Bottleneck","display_name":"LaME: Learning to Think in Latent Space for Multimodal Embedding via Information Bottleneck","publication_year":2026,"publication_date":"2026-06-11","ids":{"openalex":"https://openalex.org/W7164560501","doi":"https://doi.org/10.48550/arxiv.2606.13061"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2606.13061","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.13061","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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2606.13061","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5138535888","display_name":"Peixi Wu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Wu, Peixi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5122931632","display_name":"B Yang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yang, Biao","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138545776","display_name":"Feipeng Ma","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ma, Feipeng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138498616","display_name":"Bosong Chai","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chai, Bosong","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138537903","display_name":"Bo Lin","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Lin, Bo","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138502374","display_name":"Wei Yuan","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yuan, Wei","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138551577","display_name":"Fan Yang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yang, Fan","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138538394","display_name":"Tingting Gao","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Gao, Tingting","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5104270736","display_name":"Hebei Li","orcid":"https://orcid.org/0000-0002-7529-6331"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Li, Hebei","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5138547216","display_name":"Xiaoyan Sun","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Sun, Xiaoyan","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":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/T11273","display_name":"Advanced Graph Neural Networks","score":0.5568000078201294,"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/T11273","display_name":"Advanced Graph Neural Networks","score":0.5568000078201294,"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":0.3068000078201294,"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":0.061400000005960464,"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/embedding","display_name":"Embedding","score":0.8086000084877014},{"id":"https://openalex.org/keywords/inference","display_name":"Inference","score":0.616100013256073},{"id":"https://openalex.org/keywords/discriminative-model","display_name":"Discriminative model","score":0.5856999754905701},{"id":"https://openalex.org/keywords/pipeline","display_name":"Pipeline (software)","score":0.5504999756813049},{"id":"https://openalex.org/keywords/bottleneck","display_name":"Bottleneck","score":0.49239999055862427},{"id":"https://openalex.org/keywords/encoding","display_name":"Encoding (memory)","score":0.4514999985694885},{"id":"https://openalex.org/keywords/space","display_name":"Space (punctuation)","score":0.3357999920845032}],"concepts":[{"id":"https://openalex.org/C41608201","wikidata":"https://www.wikidata.org/wiki/Q980509","display_name":"Embedding","level":2,"score":0.8086000084877014},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6868000030517578},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.616100013256073},{"id":"https://openalex.org/C97931131","wikidata":"https://www.wikidata.org/wiki/Q5282087","display_name":"Discriminative model","level":2,"score":0.5856999754905701},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5839999914169312},{"id":"https://openalex.org/C43521106","wikidata":"https://www.wikidata.org/wiki/Q2165493","display_name":"Pipeline (software)","level":2,"score":0.5504999756813049},{"id":"https://openalex.org/C2780513914","wikidata":"https://www.wikidata.org/wiki/Q18210350","display_name":"Bottleneck","level":2,"score":0.49239999055862427},{"id":"https://openalex.org/C125411270","wikidata":"https://www.wikidata.org/wiki/Q18653","display_name":"Encoding (memory)","level":2,"score":0.4514999985694885},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3935000002384186},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.3686000108718872},{"id":"https://openalex.org/C2778572836","wikidata":"https://www.wikidata.org/wiki/Q380933","display_name":"Space (punctuation)","level":2,"score":0.3357999920845032},{"id":"https://openalex.org/C159877910","wikidata":"https://www.wikidata.org/wiki/Q2202883","display_name":"Autoregressive model","level":2,"score":0.29919999837875366},{"id":"https://openalex.org/C2164484","wikidata":"https://www.wikidata.org/wiki/Q5170150","display_name":"Core (optical fiber)","level":2,"score":0.2969000041484833},{"id":"https://openalex.org/C20162079","wikidata":"https://www.wikidata.org/wiki/Q1151406","display_name":"Case-based reasoning","level":2,"score":0.29019999504089355},{"id":"https://openalex.org/C195344581","wikidata":"https://www.wikidata.org/wiki/Q2555318","display_name":"Automated reasoning","level":2,"score":0.2867000102996826},{"id":"https://openalex.org/C43126263","wikidata":"https://www.wikidata.org/wiki/Q128751","display_name":"Source code","level":2,"score":0.2687999904155731},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.26600000262260437},{"id":"https://openalex.org/C81669768","wikidata":"https://www.wikidata.org/wiki/Q2359161","display_name":"Precision and recall","level":2,"score":0.25429999828338623},{"id":"https://openalex.org/C185429906","wikidata":"https://www.wikidata.org/wiki/Q1130160","display_name":"Estimator","level":2,"score":0.2531000077724457},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.25060001015663147}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2606.13061","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.13061","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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2606.13061","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.13061","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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"sustainable_development_goals":[{"score":0.7046633958816528,"display_name":"Reduced inequalities","id":"https://metadata.un.org/sdg/10"}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Reasoning-driven":[0],"universal":[1],"multimodal":[2],"embedding":[3,14,49,79,186],"has":[4],"advanced":[5],"rapidly":[6],"by":[7],"introducing":[8],"Chain-of-Thought":[9],"(CoT)":[10],"reasoning":[11,36,73,80,100,119],"into":[12],"the":[13,17,69],"pipeline.":[15],"Despite":[16],"strong":[18],"performance":[19,50],"across":[20],"both":[21],"general":[22],"and":[23,47,76,136,153,175],"complex":[24],"tasks,":[25],"this":[26,87],"paradigm":[27],"suffers":[28],"from":[29,133],"two":[30,126],"core":[31],"limitations:":[32],"(i)":[33],"autoregressive":[34,134],"CoT":[35,55,68,140,173],"incurs":[37],"high":[38],"computational":[39],"cost,":[40],"making":[41,58],"it":[42],"impractical":[43],"for":[44,74],"low-latency":[45],"retrieval;":[46],"(ii)":[48],"is":[51,189],"heavily":[52],"coupled":[53],"with":[54,181],"annotation":[56],"quality,":[57],"large-scale":[59],"training":[60,145],"unreliable.":[61],"These":[62],"raise":[63],"fundamental":[64],"questions:":[65],"Is":[66],"textual":[67],"optimal":[70],"form":[71],"of":[72],"embedding,":[75],"can":[77],"effective":[78],"be":[81],"accomplished":[82],"in":[83],"latent":[84,99,179],"space?":[85],"To":[86],"end,":[88],"we":[89],"propose":[90],"LaME":[91,107,157],"(Latent":[92],"Reasoning":[93],"Multimodal":[94],"Embedding),":[95],"which":[96],"formulates":[97],"embedding-oriented":[98],"as":[101,113],"a":[102,114,121,143],"weakly":[103],"supervised":[104],"information":[105],"bottleneck.":[106],"employs":[108],"K":[109],"learnable":[110],"reason":[111],"tokens":[112],"fixed-capacity":[115],"bottleneck,":[116],"completing":[117],"all":[118],"within":[120],"single":[122],"forward":[123],"pass.":[124],"The":[125],"weak":[127],"supervision":[128],"signals":[129],"structurally":[130],"decouple":[131],"contrastive":[132],"objectives":[135],"eliminate":[137],"dependence":[138],"on":[139,151],"annotations,":[141],"while":[142,166],"two-stage":[144],"pipeline":[146],"ensures":[147],"stable":[148],"convergence.":[149],"Experiments":[150],"MMEB-v2":[152],"MRMR":[154],"show":[155],"that":[156],"achieves":[158],"competitive":[159],"performance,":[160],"surpassing":[161],"some":[162],"explicit":[163,172],"CoT-based":[164],"models,":[165],"delivering":[167],"60x":[168],"faster":[169,177],"inference":[170],"than":[171,178],"methods":[174],"2x":[176],"baselines":[180],"throughput":[182],"comparable":[183],"to":[184],"discriminative":[185],"models.":[187],"Code":[188],"available":[190],"at":[191],"https://github.com/PeppaWu/LaME.":[192]},"counts_by_year":[],"updated_date":"2026-07-17T05:52:16.776730","created_date":"2026-06-13T00:00:00"}