{"id":"https://openalex.org/W7164124619","doi":"https://doi.org/10.48550/arxiv.2606.11190","title":"When to Align, When to Predict: A Phase Diagram for Multimodal Learning","display_name":"When to Align, When to Predict: A Phase Diagram for Multimodal Learning","publication_year":2026,"publication_date":"2026-06-09","ids":{"openalex":"https://openalex.org/W7164124619","doi":"https://doi.org/10.48550/arxiv.2606.11190"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2606.11190","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.11190","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":"Preprint"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2606.11190","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5103305894","display_name":"Ilay Kamai","orcid":"https://orcid.org/0009-0008-5080-496X"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Kamai, Ilay","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5031116795","display_name":"Hugues Van Assel","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Van Assel, Hugues","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5138324672","display_name":"Aviv Regev","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Regev, Aviv","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5136407649","display_name":"Hagai B. Perets","orcid":"https://orcid.org/0000-0002-5004-199X"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Perets, Hagai B.","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5138348812","display_name":"Randall Balestriero","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Balestriero, Randall","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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.43320000171661377,"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.43320000171661377,"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/T10775","display_name":"Generative Adversarial Networks and Image Synthesis","score":0.12389999628067017,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.12099999934434891,"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/representation","display_name":"Representation (politics)","score":0.6499999761581421},{"id":"https://openalex.org/keywords/diagram","display_name":"Diagram","score":0.5130000114440918},{"id":"https://openalex.org/keywords/nonlinear-system","display_name":"Nonlinear system","score":0.4399999976158142},{"id":"https://openalex.org/keywords/training-set","display_name":"Training set","score":0.41290000081062317},{"id":"https://openalex.org/keywords/code","display_name":"Code (set theory)","score":0.39640000462532043},{"id":"https://openalex.org/keywords/structured-prediction","display_name":"Structured prediction","score":0.38119998574256897},{"id":"https://openalex.org/keywords/component","display_name":"Component (thermodynamics)","score":0.3709999918937683}],"concepts":[{"id":"https://openalex.org/C2776359362","wikidata":"https://www.wikidata.org/wiki/Q2145286","display_name":"Representation (politics)","level":3,"score":0.6499999761581421},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6485000252723694},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.5461000204086304},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.5228999853134155},{"id":"https://openalex.org/C186399060","wikidata":"https://www.wikidata.org/wiki/Q959962","display_name":"Diagram","level":2,"score":0.5130000114440918},{"id":"https://openalex.org/C158622935","wikidata":"https://www.wikidata.org/wiki/Q660848","display_name":"Nonlinear system","level":2,"score":0.4399999976158142},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.41290000081062317},{"id":"https://openalex.org/C2776760102","wikidata":"https://www.wikidata.org/wiki/Q5139990","display_name":"Code (set theory)","level":3,"score":0.39640000462532043},{"id":"https://openalex.org/C22367795","wikidata":"https://www.wikidata.org/wiki/Q7625208","display_name":"Structured prediction","level":2,"score":0.38119998574256897},{"id":"https://openalex.org/C168167062","wikidata":"https://www.wikidata.org/wiki/Q1117970","display_name":"Component (thermodynamics)","level":2,"score":0.3709999918937683},{"id":"https://openalex.org/C2780226545","wikidata":"https://www.wikidata.org/wiki/Q6888030","display_name":"Modality (human\u2013computer interaction)","level":2,"score":0.3662000000476837},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.34310001134872437},{"id":"https://openalex.org/C2780660688","wikidata":"https://www.wikidata.org/wiki/Q25052564","display_name":"Multimodal learning","level":2,"score":0.3361000120639801},{"id":"https://openalex.org/C44280652","wikidata":"https://www.wikidata.org/wiki/Q104837","display_name":"Phase (matter)","level":2,"score":0.31949999928474426},{"id":"https://openalex.org/C66782513","wikidata":"https://www.wikidata.org/wiki/Q864601","display_name":"Biomedicine","level":2,"score":0.31540000438690186},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.289900004863739},{"id":"https://openalex.org/C2780910867","wikidata":"https://www.wikidata.org/wiki/Q1952416","display_name":"Multimodality","level":2,"score":0.2863999903202057},{"id":"https://openalex.org/C160920958","wikidata":"https://www.wikidata.org/wiki/Q7662746","display_name":"Synthetic data","level":2,"score":0.2849000096321106},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.27559998631477356}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2606.11190","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.11190","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":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2606.11190","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.11190","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":"Preprint"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Cross-modal":[0],"alignment":[1,102],"(CA)":[2],"and":[3,28,52,57,106,145,168,183,212],"cross-modal":[4,30,87,173,199],"prediction":[5,115,169],"(CP)":[6],"are":[7],"the":[8,66,165,188,191,195,214,224],"dominant":[9],"paradigms":[10],"for":[11,94],"multimodal":[12,135,210],"representation":[13],"learning,":[14],"yet":[15],"there":[16],"is":[17,110,118,201,226],"no":[18],"systematic":[19],"understanding":[20],"of":[21,55],"when":[22,25,29,108],"each":[23,26,104],"succeeds,":[24],"fails,":[27],"training":[31,200],"helps":[32],"at":[33,228],"all":[34],"--":[35],"a":[36,72,81,121,149,160],"gap":[37],"that":[38,76,97],"leaves":[39],"practitioners,":[40],"especially":[41],"in":[42,156,190],"scientific":[43],"domains":[44],"like":[45],"biomedicine":[46],"or":[47],"astrophysics,":[48],"with":[49,85,124],"heterogeneous":[50],"instruments":[51],"multiple":[53],"levels":[54],"organization":[56],"measurement,":[58],"unable":[59],"to":[60,152,219,222],"diagnose":[61,208],"why":[62],"standard":[63],"methods":[64],"underperform":[65],"best":[67],"single":[68],"modality.":[69],"We":[70,147],"develop":[71],"unified":[73],"linear":[74],"framework":[75,205],"addresses":[77],"both":[78,95],"questions.":[79],"Under":[80],"spiked":[82],"signal-plus-noise":[83],"model":[84],"structured":[86],"nuisance":[88,109],"correlation,":[89],"we":[90],"derive":[91],"separation":[92],"ratios":[93],"objectives":[96],"expose":[98],"complementary":[99],"failure":[100],"modes:":[101],"whitens":[103],"modality":[105],"fails":[107],"strongly":[111],"correlated":[112],"across":[113],"views;":[114],"encodes":[116],"whatever":[117],"cross-predictable":[119],"through":[120],"one-sided":[122],"whitening,":[123],"recovery":[125],"governed":[126],"by":[127],"source-modality":[128],"quality.":[129],"The":[130],"resulting":[131],"phase":[132],"diagram":[133,158],"partitions":[134],"problems":[136],"into":[137],"four":[138],"regimes:":[139],"Both,":[140],"CA":[141],"only,":[142,144],"CP":[143],"Neither.":[146],"present":[148],"data-driven":[150],"procedure":[151],"locate":[153],"real-world":[154],"datasets":[155],"this":[157],"using":[159],"small":[161],"labeled":[162],"subsample,":[163],"identifying":[164],"preferred":[166],"objective":[167,216],"direction":[170],"before":[171,217],"any":[172],"training.":[174,220],"Experiments":[175],"on":[176],"synthetic":[177],"data,":[178],"stereo-vision":[179],"benchmarks,":[180],"image-caption":[181],"pairs,":[182],"real":[184],"astrophysical":[185],"data":[186],"validate":[187],"predictions":[189],"nonlinear":[192],"regime,":[193],"including":[194],"Neither":[196],"regime":[197],"where":[198],"actively":[202],"harmful.":[203],"Our":[204],"lets":[206],"practitioners":[207],"their":[209],"problem":[211],"choose":[213],"right":[215],"committing":[218],"Code":[221],"reproduce":[223],"results":[225],"available":[227],"https://github.com/IlayMalinyak/mm_align_vs_pred.":[229]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-06-11T00:00:00"}