{"id":"https://openalex.org/W7112803773","doi":null,"title":"Group Symmetries in Diffusion Models: Formulation, Generalization, and Enforcement","display_name":"Group Symmetries in Diffusion Models: Formulation, Generalization, and Enforcement","publication_year":2025,"publication_date":"2025-01-01","ids":{"openalex":"https://openalex.org/W7112803773"},"language":"en","primary_location":{"id":"pmh:oai:dash.harvard.edu:1/42719603","doi":null,"is_oa":true,"landing_page_url":"https://dash.harvard.edu/handle/1/42719603","pdf_url":"https://dash.harvard.edu/bitstreams/a2e51691-b110-4672-a481-b2a7429ad557/download","source":{"id":"https://openalex.org/S4306401540","display_name":"Digital Access to Scholarship at Harvard (DASH) (Harvard University)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I136199984","host_organization_name":"Harvard University","host_organization_lineage":["https://openalex.org/I136199984"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Thesis or Dissertation"},"volume":null,"type":"dissertation","is_traditional":true,"indexed_in":[],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://dash.harvard.edu/bitstreams/a2e51691-b110-4672-a481-b2a7429ad557/download","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":null,"display_name":"Bharthulwar, Siddharth","orcid":null,"observed_orcids":[]},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Bharthulwar, Siddharth","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":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/T10775","display_name":"Generative Adversarial Networks and Image Synthesis","score":0.9487000107765198,"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/T10775","display_name":"Generative Adversarial Networks and Image Synthesis","score":0.9487000107765198,"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/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.017400000244379044,"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/T11206","display_name":"Model Reduction and Neural Networks","score":0.008799999952316284,"subfield":{"id":"https://openalex.org/subfields/3109","display_name":"Statistical and Nonlinear Physics"},"field":{"id":"https://openalex.org/fields/31","display_name":"Physics and Astronomy"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}}],"keywords":[{"id":"https://openalex.org/keywords/diffusion-models","display_name":"diffusion models","score":0.8320000171661377},{"id":"https://openalex.org/keywords/group-symmetry","display_name":"group symmetry","score":0.972000002861023},{"id":"https://openalex.org/keywords/equivariance","display_name":"equivariance","score":0.781000018119812},{"id":"https://openalex.org/keywords/partial-observability","display_name":"partial observability","score":0.40299999713897705},{"id":"https://openalex.org/keywords/neural-tangent-kernel","display_name":"neural tangent kernel","score":0.7250000238418579},{"id":"https://openalex.org/keywords/structured-data-generation","display_name":"structured data generation","score":0.578000009059906},{"id":"https://openalex.org/keywords/data-augmentation","display_name":"data augmentation","score":0.39399999380111694}],"concepts":[{"id":"https://openalex.org/C2779886137","wikidata":"https://www.wikidata.org/wiki/Q21030012","display_name":"Symmetry (geometry)","level":2,"score":0.5182999968528748},{"id":"https://openalex.org/C177148314","wikidata":"https://www.wikidata.org/wiki/Q170084","display_name":"Generalization","level":2,"score":0.5034000277519226},{"id":"https://openalex.org/C132459708","wikidata":"https://www.wikidata.org/wiki/Q744069","display_name":"Extrapolation","level":2,"score":0.46709999442100525},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.46209999918937683},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.45829999446868896},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.45239999890327454},{"id":"https://openalex.org/C167966045","wikidata":"https://www.wikidata.org/wiki/Q5532625","display_name":"Generative model","level":3,"score":0.4453999996185303},{"id":"https://openalex.org/C96469262","wikidata":"https://www.wikidata.org/wiki/Q1324364","display_name":"Homogeneous space","level":2,"score":0.44200000166893005},{"id":"https://openalex.org/C74193536","wikidata":"https://www.wikidata.org/wiki/Q574844","display_name":"Kernel (algebra)","level":2,"score":0.40369999408721924},{"id":"https://openalex.org/C101454708","wikidata":"https://www.wikidata.org/wiki/Q17106019","display_name":"Standard Model (mathematical formulation)","level":3,"score":0.3959999978542328},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.38510000705718994},{"id":"https://openalex.org/C2781311116","wikidata":"https://www.wikidata.org/wiki/Q83306","display_name":"Group (periodic table)","level":2,"score":0.3634999990463257},{"id":"https://openalex.org/C69357855","wikidata":"https://www.wikidata.org/wiki/Q163214","display_name":"Diffusion","level":2,"score":0.35440000891685486},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.3206000030040741},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.31709998846054077},{"id":"https://openalex.org/C44306375","wikidata":"https://www.wikidata.org/wiki/Q902019","display_name":"Symmetry group","level":2,"score":0.3151000142097473},{"id":"https://openalex.org/C39890363","wikidata":"https://www.wikidata.org/wiki/Q36108","display_name":"Generative grammar","level":2,"score":0.3125},{"id":"https://openalex.org/C9652623","wikidata":"https://www.wikidata.org/wiki/Q190109","display_name":"Field (mathematics)","level":2,"score":0.3109999895095825},{"id":"https://openalex.org/C21080849","wikidata":"https://www.wikidata.org/wiki/Q13611879","display_name":"Data point","level":2,"score":0.2985000014305115},{"id":"https://openalex.org/C138187205","wikidata":"https://www.wikidata.org/wiki/Q131251","display_name":"Tangent","level":2,"score":0.29679998755455017},{"id":"https://openalex.org/C22019652","wikidata":"https://www.wikidata.org/wiki/Q331309","display_name":"Overfitting","level":3,"score":0.28760001063346863},{"id":"https://openalex.org/C106131492","wikidata":"https://www.wikidata.org/wiki/Q3072260","display_name":"Filter (signal processing)","level":2,"score":0.2816999852657318},{"id":"https://openalex.org/C98045186","wikidata":"https://www.wikidata.org/wiki/Q205663","display_name":"Process (computing)","level":2,"score":0.27480000257492065},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.26330000162124634},{"id":"https://openalex.org/C12426560","wikidata":"https://www.wikidata.org/wiki/Q189569","display_name":"Basis (linear algebra)","level":2,"score":0.2565999925136566},{"id":"https://openalex.org/C190502265","wikidata":"https://www.wikidata.org/wiki/Q17069496","display_name":"MNIST database","level":3,"score":0.2540000081062317}],"mesh":[],"locations_count":1,"locations":[{"id":"pmh:oai:dash.harvard.edu:1/42719603","doi":null,"is_oa":true,"landing_page_url":"https://dash.harvard.edu/handle/1/42719603","pdf_url":"https://dash.harvard.edu/bitstreams/a2e51691-b110-4672-a481-b2a7429ad557/download","source":{"id":"https://openalex.org/S4306401540","display_name":"Digital Access to Scholarship at Harvard (DASH) (Harvard University)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I136199984","host_organization_name":"Harvard University","host_organization_lineage":["https://openalex.org/I136199984"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Thesis or Dissertation"}],"best_oa_location":{"id":"pmh:oai:dash.harvard.edu:1/42719603","doi":null,"is_oa":true,"landing_page_url":"https://dash.harvard.edu/handle/1/42719603","pdf_url":"https://dash.harvard.edu/bitstreams/a2e51691-b110-4672-a481-b2a7429ad557/download","source":{"id":"https://openalex.org/S4306401540","display_name":"Digital Access to Scholarship at Harvard (DASH) (Harvard University)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I136199984","host_organization_name":"Harvard University","host_organization_lineage":["https://openalex.org/I136199984"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Thesis or Dissertation"},"sustainable_development_goals":[],"study_designs":[],"awards":[],"funders":[],"has_content":{"grobid_xml":true,"pdf":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7112803773.pdf","grobid_xml":"https://content.openalex.org/works/W7112803773.grobid-xml"},"referenced_works_count":17,"referenced_works":["https://openalex.org/W1959608418","https://openalex.org/W2409550820","https://openalex.org/W3013559175","https://openalex.org/W3122887982","https://openalex.org/W3157286395","https://openalex.org/W3171974924","https://openalex.org/W3199237569","https://openalex.org/W4221159371","https://openalex.org/W4224035735","https://openalex.org/W4281398962","https://openalex.org/W4293763266","https://openalex.org/W4303647933","https://openalex.org/W4388918580","https://openalex.org/W4390872297","https://openalex.org/W4400644035","https://openalex.org/W4415796698","https://openalex.org/W7133208595"],"related_works":["https://openalex.org/W7124729231","https://openalex.org/W4403344759","https://openalex.org/W4415796424","https://openalex.org/W4417298401","https://openalex.org/W4416046183","https://openalex.org/W4415308228","https://openalex.org/W4403885729","https://openalex.org/W7155532689","https://openalex.org/W7129303901","https://openalex.org/W7133233145"],"abstract_inverted_index":{"Group":[0],"symmetries":[1,60,90],"are":[2],"fundamental":[3],"structures":[4],"in":[5,29,81,103,143,188,206],"many":[6],"real-world":[7],"datasets,":[8],"and":[9,18,56,97,132,152,175,210],"lever-":[10],"aging":[11],"them":[12],"is":[13,67],"crucial":[14],"for":[15,40,223],"building":[16],"robust":[17],"data-efficient":[19],"machine":[20],"learning":[21],"models.":[22],"Diffusion":[23],"models":[24,52,105,126,209],"have":[25],"achieved":[26],"state-of-the-art":[27],"performance":[28],"generative":[30,225],"tasks":[31],"but":[32],"typically":[33],"rely":[34],"on":[35,128,170,227],"standard":[36,48,85,125,189,207],"neural":[37],"network":[38],"architectures":[39],"score":[41,101,140],"estimation.":[42],"This":[43,200],"thesis":[44],"investigates":[45],"whether":[46],"such":[47],"configurations":[49],"enable":[50],"diffusion":[51,104,208],"to":[53,88,135,164,196],"implicitly":[54],"learn":[55],"generalize":[57,89,136],"underlying":[58],"data":[59,66,93,130,178,198],"purely":[61],"from":[62,72],"examples,":[63],"particularly":[64],"when":[65],"partially":[68],"observed.":[69],"Drawing":[70],"motivation":[71],"Neural":[73],"Tangent":[74],"Kernel":[75],"(NTK)":[76],"theory,":[77],"which":[78],"suggests":[79],"limitations":[80],"the":[82,161,172,212],"ability":[83],"of":[84,124,214],"supervised":[86],"networks":[87,102],"beyond":[91],"local":[92],"structure,":[94],"we":[95,119,150],"hypothesize":[96],"empirically":[98],"demonstrate":[99],"that":[100,159,180],"exhibit":[106],"similar":[107],"constraints.":[108],"Using":[109],"a":[110,121,154,203],"2D":[111],"toy":[112,173],"dataset":[113,174],"with":[114],"inherent":[115],"SO(2)":[116],"rotational":[117],"symmetry,":[118,137],"show":[120],"consistent":[122,193],"failure":[123],"trained":[127],"incomplete":[129],"(interpolation":[131],"extrapolation":[133],"settings)":[134],"exhibiting":[138],"significant":[139],"field":[141],"distortions":[142],"unobserved":[144],"regions.":[145],"To":[146],"address":[147],"this":[148,181],"limitation,":[149],"propose":[151],"evaluate":[153],"novel":[155],"per-timestep":[156],"symmetry":[157,185],"loss":[158,182],"regularizes":[160],"denoising":[162],"process":[163],"encourage":[165],"approximate":[166],"equivariance.":[167],"Empirical":[168],"results":[169,194],"both":[171],"higher-dimensional":[176],"MNIST":[177],"confirm":[179],"significantly":[183],"enhances":[184],"generalization":[186],"even":[187],"architectures,":[190],"yielding":[191],"geometrically":[192],"comparable":[195],"extensive":[197],"augmentation.":[199],"work":[201],"highlights":[202],"critical":[204],"limitation":[205],"underscores":[211],"importance":[213],"incorporating":[215],"explicit":[216],"geometric":[217],"biases,":[218],"via":[219],"architecture":[220],"or":[221],"regularization,":[222],"reliable":[224],"modeling":[226],"structured":[228],"data.":[229]},"counts_by_year":[],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2025-12-10T00:00:00"}