{"id":"https://openalex.org/W4415288781","doi":"https://doi.org/10.1109/iccv51701.2025.01555","title":"FreeFlux: Understanding and Exploiting Layer-Specific Roles in RoPE-Based MMDiT for Versatile Image Editing","display_name":"FreeFlux: Understanding and Exploiting Layer-Specific Roles in RoPE-Based MMDiT for Versatile Image Editing","publication_year":2025,"publication_date":"2025-10-19","ids":{"openalex":"https://openalex.org/W4415288781","doi":"https://doi.org/10.1109/iccv51701.2025.01555"},"language":"en","primary_location":{"id":"doi:10.1109/iccv51701.2025.01555","is_oa":false,"landing_page_url":"https://doi.org/10.1109/iccv51701.2025.01555","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE/CVF International Conference on Computer Vision (ICCV)","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["arxiv","crossref","datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://arxiv.org/pdf/2503.16153","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5006214571","display_name":"Tianyi Wei","orcid":"https://orcid.org/0000-0002-7976-8439"},"institutions":[{"id":"https://openalex.org/I172675005","display_name":"Nanyang Technological University","ror":"https://ror.org/02e7b5302","country_code":"SG","type":"education","lineage":["https://openalex.org/I172675005"]}],"countries":["SG"],"is_corresponding":false,"raw_author_name":"Tianyi Wei","raw_affiliation_strings":["Nanyang Technological University,S-Lab"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Nanyang Technological University,S-Lab","institution_ids":["https://openalex.org/I172675005"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5101879769","display_name":"Yifan Zhou","orcid":"https://orcid.org/0009-0001-6288-6113"},"institutions":[{"id":"https://openalex.org/I172675005","display_name":"Nanyang Technological University","ror":"https://ror.org/02e7b5302","country_code":"SG","type":"education","lineage":["https://openalex.org/I172675005"]}],"countries":["SG"],"is_corresponding":false,"raw_author_name":"Yifan Zhou","raw_affiliation_strings":["Nanyang Technological University,S-Lab"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Nanyang Technological University,S-Lab","institution_ids":["https://openalex.org/I172675005"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100364587","display_name":"Dongdong Chen","orcid":"https://orcid.org/0000-0002-4642-4373"},"institutions":[{"id":"https://openalex.org/I1290206253","display_name":"Microsoft (United States)","ror":"https://ror.org/00d0nc645","country_code":"US","type":"company","lineage":["https://openalex.org/I1290206253"]},{"id":"https://openalex.org/I4210164937","display_name":"Microsoft Research (United Kingdom)","ror":"https://ror.org/05k87vq12","country_code":"GB","type":"company","lineage":["https://openalex.org/I1290206253","https://openalex.org/I4210164937"]}],"countries":["GB","US"],"is_corresponding":false,"raw_author_name":"Dongdong Chen","raw_affiliation_strings":["Microsoft GenAI"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Microsoft GenAI","institution_ids":["https://openalex.org/I1290206253","https://openalex.org/I4210164937"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5108927117","display_name":"Xingang Pan","orcid":null},"institutions":[{"id":"https://openalex.org/I172675005","display_name":"Nanyang Technological University","ror":"https://ror.org/02e7b5302","country_code":"SG","type":"education","lineage":["https://openalex.org/I172675005"]}],"countries":["SG"],"is_corresponding":false,"raw_author_name":"Xingang Pan","raw_affiliation_strings":["Nanyang Technological University,S-Lab"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":"Nanyang Technological University,S-Lab","institution_ids":["https://openalex.org/I172675005"]}]}],"institutions":[],"countries_distinct_count":3,"institutions_distinct_count":3,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":0.0,"has_fulltext":false,"cited_by_count":0,"citation_normalized_percentile":{"value":0.35591709,"is_in_top_1_percent":false,"is_in_top_10_percent":false},"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"16745","last_page":"16754"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10715","display_name":"Distributed and Parallel Computing Systems","score":0.8855999708175659,"subfield":{"id":"https://openalex.org/subfields/1705","display_name":"Computer Networks and Communications"},"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/T10715","display_name":"Distributed and Parallel Computing Systems","score":0.8855999708175659,"subfield":{"id":"https://openalex.org/subfields/1705","display_name":"Computer Networks and Communications"},"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/image-editing","display_name":"Image editing","score":0.8177000284194946},{"id":"https://openalex.org/keywords/embedding","display_name":"Embedding","score":0.7143999934196472},{"id":"https://openalex.org/keywords/video-editing","display_name":"Video editing","score":0.5501000285148621},{"id":"https://openalex.org/keywords/object","display_name":"Object (grammar)","score":0.41909998655319214},{"id":"https://openalex.org/keywords/similarity","display_name":"Similarity (geometry)","score":0.39089998602867126},{"id":"https://openalex.org/keywords/image","display_name":"Image (mathematics)","score":0.3846000134944916},{"id":"https://openalex.org/keywords/dependency","display_name":"Dependency (UML)","score":0.36550000309944153},{"id":"https://openalex.org/keywords/transformer","display_name":"Transformer","score":0.34369999170303345}],"concepts":[{"id":"https://openalex.org/C2776674983","wikidata":"https://www.wikidata.org/wiki/Q545981","display_name":"Image editing","level":3,"score":0.8177000284194946},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.8116999864578247},{"id":"https://openalex.org/C41608201","wikidata":"https://www.wikidata.org/wiki/Q980509","display_name":"Embedding","level":2,"score":0.7143999934196472},{"id":"https://openalex.org/C2780310081","wikidata":"https://www.wikidata.org/wiki/Q1154312","display_name":"Video editing","level":2,"score":0.5501000285148621},{"id":"https://openalex.org/C107457646","wikidata":"https://www.wikidata.org/wiki/Q207434","display_name":"Human\u2013computer interaction","level":1,"score":0.44519999623298645},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.426800012588501},{"id":"https://openalex.org/C2781238097","wikidata":"https://www.wikidata.org/wiki/Q175026","display_name":"Object (grammar)","level":2,"score":0.41909998655319214},{"id":"https://openalex.org/C103278499","wikidata":"https://www.wikidata.org/wiki/Q254465","display_name":"Similarity (geometry)","level":3,"score":0.39089998602867126},{"id":"https://openalex.org/C115961682","wikidata":"https://www.wikidata.org/wiki/Q860623","display_name":"Image (mathematics)","level":2,"score":0.3846000134944916},{"id":"https://openalex.org/C19768560","wikidata":"https://www.wikidata.org/wiki/Q320727","display_name":"Dependency (UML)","level":2,"score":0.36550000309944153},{"id":"https://openalex.org/C23123220","wikidata":"https://www.wikidata.org/wiki/Q816826","display_name":"Information retrieval","level":1,"score":0.3499999940395355},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.34369999170303345},{"id":"https://openalex.org/C98045186","wikidata":"https://www.wikidata.org/wiki/Q205663","display_name":"Process (computing)","level":2,"score":0.3366999924182892},{"id":"https://openalex.org/C2987933465","wikidata":"https://www.wikidata.org/wiki/Q141130","display_name":"Image manipulation","level":3,"score":0.320499986410141},{"id":"https://openalex.org/C31972630","wikidata":"https://www.wikidata.org/wiki/Q844240","display_name":"Computer vision","level":1,"score":0.3131999969482422},{"id":"https://openalex.org/C130318100","wikidata":"https://www.wikidata.org/wiki/Q2268914","display_name":"Semantic similarity","level":2,"score":0.30410000681877136},{"id":"https://openalex.org/C12713177","wikidata":"https://www.wikidata.org/wiki/Q1900281","display_name":"Perspective (graphical)","level":2,"score":0.29739999771118164},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.29280000925064087},{"id":"https://openalex.org/C87829876","wikidata":"https://www.wikidata.org/wiki/Q648877","display_name":"Post-production","level":2,"score":0.28189998865127563},{"id":"https://openalex.org/C125411270","wikidata":"https://www.wikidata.org/wiki/Q18653","display_name":"Encoding (memory)","level":2,"score":0.2777999937534332},{"id":"https://openalex.org/C171268870","wikidata":"https://www.wikidata.org/wiki/Q1486676","display_name":"GRASP","level":2,"score":0.27000001072883606},{"id":"https://openalex.org/C198082294","wikidata":"https://www.wikidata.org/wiki/Q3399648","display_name":"Position (finance)","level":2,"score":0.25540000200271606},{"id":"https://openalex.org/C3018587665","wikidata":"https://www.wikidata.org/wiki/Q7268696","display_name":"Qualitative analysis","level":3,"score":0.25519999861717224}],"mesh":[],"locations_count":3,"locations":[{"id":"doi:10.1109/iccv51701.2025.01555","is_oa":false,"landing_page_url":"https://doi.org/10.1109/iccv51701.2025.01555","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"2025 IEEE/CVF International Conference on Computer Vision (ICCV)","raw_type":"proceedings-article"},{"id":"pmh:oai:arXiv.org:2503.16153","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2503.16153","pdf_url":"https://arxiv.org/pdf/2503.16153","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,"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.2503.16153","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2503.16153","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,"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":"pmh:oai:arXiv.org:2503.16153","is_oa":true,"landing_page_url":"http://arxiv.org/abs/2503.16153","pdf_url":"https://arxiv.org/pdf/2503.16153","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,"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":[],"awards":[{"id":"https://openalex.org/G2272717301","display_name":null,"funder_award_id":"NRF-NRFF16-2024-0003","funder_id":"https://openalex.org/F4320320709","funder_display_name":"National Research Foundation Singapore"}],"funders":[{"id":"https://openalex.org/F4320320709","display_name":"National Research Foundation Singapore","ror":"https://ror.org/03cpyc314"}],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"The":[0],"integration":[1],"of":[2,22,43,139],"Rotary":[3],"Position":[4],"Embedding":[5],"(RoPE)":[6],"in":[7,87,156],"Multimodal":[8],"Diffusion":[9],"Transformer":[10],"(MMDiT)":[11],"has":[12],"significantly":[13],"enhanced":[14],"text-to-image":[15],"generation":[16,32],"quality.":[17],"However,":[18],"the":[19,39,84,137,140],"fundamental":[20],"reliance":[21],"self-attention":[23],"layers":[24],"on":[25,91,136],"positional":[26,56],"embedding":[27],"versus":[28,58],"query-key":[29],"similarity":[30],"during":[31,65],"remains":[33],"an":[34,50],"intriguing":[35],"question.":[36],"We":[37],"present":[38],"first":[40],"mechanistic":[41],"analysis":[42,68],"RoPE-based":[44,88],"MMDiT":[45],"models":[46],"(e.g.,":[47,111,117,123],"FLUX),":[48],"introducing":[49],"automated":[51],"probing":[52],"strategy":[53],"that":[54,73,102,149],"disentangles":[55],"information":[57],"content":[59,114,160],"dependencies":[60],"by":[61],"strategically":[62],"manipulating":[63],"RoPE":[64],"generation.":[66],"Our":[67],"reveals":[69],"distinct":[70],"dependency":[71],"patterns":[72],"do":[74],"not":[75],"straightforwardly":[76],"correlate":[77],"with":[78],"depth,":[79],"offering":[80],"new":[81],"insights":[82],"into":[83,106],"layer-specific":[85],"roles":[86],"MMDiT.":[89],"Based":[90],"these":[92],"findings,":[93],"we":[94,129],"propose":[95],"a":[96],"training-free,":[97],"task-specific":[98],"image":[99],"editing":[100,104,110,116,122,141],"framework":[101],"categorizes":[103],"tasks":[105],"three":[107],"types:":[108],"position-dependent":[109],"object":[112],"addition),":[113],"similarity-dependent":[115],"non-rigid":[118],"editing),":[119],"and":[120,145,161],"region-preserved":[121],"background":[124],"replacement).":[125],"For":[126],"each":[127],"type,":[128],"design":[130],"tailored":[131],"key-value":[132],"injection":[133],"strategies":[134],"based":[135],"characteristics":[138],"task.":[142],"Extensive":[143],"qualitative":[144],"quantitative":[146],"evaluations":[147],"demonstrate":[148],"our":[150],"method":[151],"outperforms":[152],"state-of-the-art":[153],"approaches,":[154],"particularly":[155],"preserving":[157],"original":[158],"semantic":[159],"achieving":[162],"seamless":[163],"modifications.":[164]},"counts_by_year":[],"updated_date":"2026-07-29T14:22:42.915294","created_date":"2025-10-17T00:00:00"}