{"id":"https://openalex.org/W7220683963","doi":"https://doi.org/10.48550/arxiv.2610.05043","title":"CI-JEPA: A Counterfactual Analysis of Latent Representations in Joint-Embedding Predictive Architectures for Self-Supervised Learning","display_name":"CI-JEPA: A Counterfactual Analysis of Latent Representations in Joint-Embedding Predictive Architectures for Self-Supervised Learning","publication_year":2026,"publication_date":"2026-10-04","ids":{"openalex":"https://openalex.org/W7220683963","doi":"https://doi.org/10.48550/arxiv.2610.05043"},"language":"en","primary_location":{"id":"pmh:oai:arXiv.org:2610.05043","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2610.05043","pdf_url":"https://arxiv.org/pdf/2610.05043","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":null,"license_id":null,"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/2610.05043","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5133765387","display_name":"Mintu Dutta","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I33586908","display_name":"Pandit Deendayal Energy University","ror":"https://ror.org/02nsv5p42","country_code":"IN","type":"education","lineage":["https://openalex.org/I33586908"]}],"countries":["IN"],"is_corresponding":false,"raw_author_name":"Dutta, Mintu","raw_affiliation_strings":["Central University , India","Department of Information and Communication Technology , Pandit Deendayal Energy University , Gandhinagar , Gujarat , India"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Central University , India","institution_ids":[]},{"raw_affiliation_string":"Department of Information and Communication Technology , Pandit Deendayal Energy University , Gandhinagar , Gujarat , India","institution_ids":["https://openalex.org/I33586908"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5032318934","display_name":"Ritesh Vyas","orcid":"https://orcid.org/0000-0002-9739-2551","observed_orcids":["https://orcid.org/0000-0002-9739-2551"]},"institutions":[{"id":"https://openalex.org/I74796645","display_name":"Birla Institute of Technology and Science, Pilani","ror":"https://ror.org/001p3jz28","country_code":"IN","type":"education","lineage":["https://openalex.org/I74796645"]}],"countries":["IN"],"is_corresponding":false,"raw_author_name":"Vyas, Ritesh","raw_affiliation_strings":["Department of Electrical and Electronics Engineering , Birla Institute of Technology & Science , Pilani , Rajasthan , India"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Department of Electrical and Electronics Engineering , Birla Institute of Technology & Science , Pilani , Rajasthan , India","institution_ids":["https://openalex.org/I74796645"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5157663722","display_name":"Mohendra Roy *","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"*, Mohendra Roy","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.8984000086784363,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.8984000086784363,"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/T10036","display_name":"Advanced Neural Network Applications","score":0.041600000113248825,"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":0.01679999940097332,"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/jepa","display_name":"JEPA","score":0.6679999828338623},{"id":"https://openalex.org/keywords/counterfactual-intervention","display_name":"counterfactual intervention","score":0.6880000233650208},{"id":"https://openalex.org/keywords/self-supervised-visual-representation-learning","display_name":"self-supervised visual representation learning","score":0.8820000290870667},{"id":"https://openalex.org/keywords/representation-robustness","display_name":"representation robustness","score":0.7749999761581421},{"id":"https://openalex.org/keywords/occlusion","display_name":"occlusion","score":0.5820000171661377},{"id":"https://openalex.org/keywords/background-blur","display_name":"background blur","score":0.7310000061988831},{"id":"https://openalex.org/keywords/linear-probing","display_name":"linear probing","score":0.6179999709129333}],"concepts":[{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.8259999752044678},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.72079998254776},{"id":"https://openalex.org/C108650721","wikidata":"https://www.wikidata.org/wiki/Q1783253","display_name":"Counterfactual thinking","level":2,"score":0.6732000112533569},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.628600001335144},{"id":"https://openalex.org/C2776359362","wikidata":"https://www.wikidata.org/wiki/Q2145286","display_name":"Representation (politics)","level":3,"score":0.5356000065803528},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.5151000022888184},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.44699999690055847},{"id":"https://openalex.org/C56461940","wikidata":"https://www.wikidata.org/wiki/Q970687","display_name":"Eye tracking","level":2,"score":0.4165000021457672},{"id":"https://openalex.org/C2776401178","wikidata":"https://www.wikidata.org/wiki/Q12050496","display_name":"Feature (linguistics)","level":2,"score":0.3862000107765198},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.31610000133514404},{"id":"https://openalex.org/C184337299","wikidata":"https://www.wikidata.org/wiki/Q1437428","display_name":"Semantics (computer science)","level":2,"score":0.30970001220703125},{"id":"https://openalex.org/C36464697","wikidata":"https://www.wikidata.org/wiki/Q451553","display_name":"Visualization","level":2,"score":0.3093000054359436},{"id":"https://openalex.org/C115961682","wikidata":"https://www.wikidata.org/wiki/Q860623","display_name":"Image (mathematics)","level":2,"score":0.28929999470710754},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.28529998660087585},{"id":"https://openalex.org/C59404180","wikidata":"https://www.wikidata.org/wiki/Q17013334","display_name":"Feature learning","level":2,"score":0.2676999866962433},{"id":"https://openalex.org/C52622490","wikidata":"https://www.wikidata.org/wiki/Q1026626","display_name":"Feature extraction","level":2,"score":0.26260000467300415},{"id":"https://openalex.org/C97256817","wikidata":"https://www.wikidata.org/wiki/Q1462316","display_name":"Spurious relationship","level":2,"score":0.2606000006198883},{"id":"https://openalex.org/C75294576","wikidata":"https://www.wikidata.org/wiki/Q5165192","display_name":"Contextual image classification","level":3,"score":0.25209999084472656}],"mesh":[],"locations_count":2,"locations":[{"id":"pmh:oai:arXiv.org:2610.05043","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2610.05043","pdf_url":"https://arxiv.org/pdf/2610.05043","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":null,"license_id":null,"version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"text"},{"id":"doi:10.48550/arxiv.2610.05043","doi":"https://doi.org/10.48550/arxiv.2610.05043","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2610.05043","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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"pmh:oai:arXiv.org:2610.05043","doi":null,"is_oa":true,"landing_page_url":"https://arxiv.org/abs/2610.05043","pdf_url":"https://arxiv.org/pdf/2610.05043","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":null,"license_id":null,"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/W7220683963.pdf","grobid_xml":"https://content.openalex.org/works/W7220683963.grobid-xml"},"referenced_works_count":4,"referenced_works":["https://openalex.org/W2529194139","https://openalex.org/W3036438747","https://openalex.org/W4297411720","https://openalex.org/W4386076428"],"related_works":["https://openalex.org/W4415799137","https://openalex.org/W4416598848","https://openalex.org/W4386076428","https://openalex.org/W7160439851","https://openalex.org/W7126524120","https://openalex.org/W7220779406","https://openalex.org/W7167584582","https://openalex.org/W4406925504","https://openalex.org/W7164505169","https://openalex.org/W7054840859"],"abstract_inverted_index":{"Self-supervised":[0],"visual":[1,36,208],"representation":[2,11,50,67,136],"learning":[3],"learns":[4,46],"useful":[5],"features":[6],"without":[7],"manual":[8],"annotations":[9],"during":[10],"training.":[12],"The":[13,132,169],"image-based":[14],"joint-embedding":[15],"predictive":[16],"architecture":[17],"(I-JEPA)":[18],"predicts":[19],"latent":[20],"representations":[21],"of":[22,112,128],"masked":[23],"image":[24,60,185],"regions,":[25],"but":[26],"its":[27],"objective":[28],"does":[29,175],"not":[30,176,199],"explicitly":[31],"model":[32],"responses":[33,73],"to":[34,47,74,79,206],"specified":[35],"interventions.":[37,101],"We":[38,65],"introduce":[39],"CI-JEPA,":[40],"a":[41,62,89,108,126,188],"counterfactual":[42],"intervention-aware":[43],"extension":[44],"that":[45],"predict":[48],"the":[49,119,123,161,181],"change":[51],"$\u0394Z":[52],"=":[53],"Z_{\\mathrm{CF}}":[54],"-":[55],"Z$":[56],"between":[57],"an":[58],"original":[59],"and":[61,93,96,122,147,174],"modified":[63],"counterpart.":[64],"assess":[66],"robustness":[68,179,205],"through":[69],"selective":[70],"sensitivity:":[71],"stronger":[72],"task-relevant":[75],"semantic":[76,91,158],"changes":[77,137,193],"than":[78,165],"nuisance":[80,100,167],"changes.":[81,209],"Experiments":[82],"on":[83],"Flowers102":[84],"use":[85],"flower-center":[86],"occlusion":[87],"as":[88,98],"candidate":[90,99],"intervention":[92],"background":[94,145,150],"blur":[95],"tint":[97],"With":[102],"frozen-encoder":[103],"linear":[104],"probing,":[105],"CI-JEPA":[106],"achieves":[107],"best":[109],"validation":[110],"accuracy":[111,170],"78.14\\%,":[113],"compared":[114],"with":[115,156],"77.55\\%":[116],"for":[117,140,144,149,160,190],"both":[118],"pretrained":[120],"ViT-B/16":[121],"I-JEPA":[124],"baseline,":[125],"gain":[127],"0.59":[129],"percentage":[130],"points.":[131],"reported":[133],"mean":[134],"$L_2$":[135],"are":[138],"4.42":[139],"center":[141],"occlusion,":[142],"3.48":[143],"tint,":[146],"2.83":[148],"blur.":[151],"This":[152],"ordering":[153],"is":[154,172],"consistent":[155],"relative":[157],"selectivity":[159],"evaluated":[162],"interventions,":[163],"rather":[164],"complete":[166],"invariance.":[168],"comparison":[171],"complementary":[173],"establish":[177,200],"improved":[178],"over":[180],"baselines.":[182],"These":[183],"controlled":[184],"modifications":[186],"provide":[187],"framework":[189],"studying":[191],"intervention-induced":[192],"in":[194],"JEPA":[195],"representations;":[196],"they":[197],"do":[198],"causal":[201],"feature":[202],"discovery":[203],"or":[204],"all":[207]},"counts_by_year":[],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2026-10-07T00:00:00"}