{"id":"https://openalex.org/W7131389935","doi":"https://doi.org/10.48550/arxiv.2602.19031","title":"SKYLIGHT: A Scalable Hundred-Channel 3D Photonic In-Memory Tensor Core Architecture for Real-time AI Inference","display_name":"SKYLIGHT: A Scalable Hundred-Channel 3D Photonic In-Memory Tensor Core Architecture for Real-time AI Inference","publication_year":2026,"publication_date":"2026-02-22","ids":{"openalex":"https://openalex.org/W7131389935","doi":"https://doi.org/10.48550/arxiv.2602.19031"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2602.19031","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2602.19031","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":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.2602.19031","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5126835463","display_name":"Meng Zhang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Zhang, Meng","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5126782868","display_name":"Ziang Yin","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Yin, Ziang","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5093925841","display_name":"Nicholas Gangi","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Gangi, Nicholas","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5126841481","display_name":"Alexander Chen","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Chen, Alexander","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5126723366","display_name":"Brett Bamfo","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Bamfo, Brett","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5126345316","display_name":"Tianle Xu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xu, Tianle","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5126823802","display_name":"Jiaqi Gu","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Gu, Jiaqi","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5126813190","display_name":"Zhaoran Rena Huang","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Huang, Zhaoran Rena","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/T12611","display_name":"Neural Networks and Reservoir Computing","score":0.9351999759674072,"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/T12611","display_name":"Neural Networks and Reservoir Computing","score":0.9351999759674072,"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/T10299","display_name":"Photonic and Optical Devices","score":0.023900000378489494,"subfield":{"id":"https://openalex.org/subfields/2208","display_name":"Electrical and Electronic Engineering"},"field":{"id":"https://openalex.org/fields/22","display_name":"Engineering"},"domain":{"id":"https://openalex.org/domains/3","display_name":"Physical Sciences"}},{"id":"https://openalex.org/T11996","display_name":"Random lasers and scattering media","score":0.003700000001117587,"subfield":{"id":"https://openalex.org/subfields/3102","display_name":"Acoustics and Ultrasonics"},"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/photonics","display_name":"Photonics","score":0.6431000232696533},{"id":"https://openalex.org/keywords/scalability","display_name":"Scalability","score":0.6396999955177307},{"id":"https://openalex.org/keywords/softmax-function","display_name":"Softmax function","score":0.5307999849319458},{"id":"https://openalex.org/keywords/robustness","display_name":"Robustness (evolution)","score":0.4465000033378601},{"id":"https://openalex.org/keywords/skylight","display_name":"Skylight","score":0.3709000051021576},{"id":"https://openalex.org/keywords/multiplexing","display_name":"Multiplexing","score":0.35359999537467957}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7170000076293945},{"id":"https://openalex.org/C20788544","wikidata":"https://www.wikidata.org/wiki/Q467054","display_name":"Photonics","level":2,"score":0.6431000232696533},{"id":"https://openalex.org/C48044578","wikidata":"https://www.wikidata.org/wiki/Q727490","display_name":"Scalability","level":2,"score":0.6396999955177307},{"id":"https://openalex.org/C188441871","wikidata":"https://www.wikidata.org/wiki/Q7554146","display_name":"Softmax function","level":3,"score":0.5307999849319458},{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.4465000033378601},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.42100000381469727},{"id":"https://openalex.org/C2780574069","wikidata":"https://www.wikidata.org/wiki/Q2712955","display_name":"Skylight","level":2,"score":0.3709000051021576},{"id":"https://openalex.org/C19275194","wikidata":"https://www.wikidata.org/wiki/Q222903","display_name":"Multiplexing","level":2,"score":0.35359999537467957},{"id":"https://openalex.org/C118524514","wikidata":"https://www.wikidata.org/wiki/Q173212","display_name":"Computer architecture","level":1,"score":0.3375999927520752},{"id":"https://openalex.org/C24326235","wikidata":"https://www.wikidata.org/wiki/Q126095","display_name":"Electronic engineering","level":1,"score":0.3287999927997589},{"id":"https://openalex.org/C104267543","wikidata":"https://www.wikidata.org/wiki/Q208163","display_name":"Signal processing","level":3,"score":0.3160000145435333},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.30379998683929443},{"id":"https://openalex.org/C2776214188","wikidata":"https://www.wikidata.org/wiki/Q408386","display_name":"Inference","level":2,"score":0.30379998683929443},{"id":"https://openalex.org/C459310","wikidata":"https://www.wikidata.org/wiki/Q117801","display_name":"Computational science","level":1,"score":0.3019999861717224},{"id":"https://openalex.org/C113775141","wikidata":"https://www.wikidata.org/wiki/Q428691","display_name":"Computer engineering","level":1,"score":0.2809999883174896},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.2777999937534332},{"id":"https://openalex.org/C22799297","wikidata":"https://www.wikidata.org/wiki/Q523846","display_name":"Photonic integrated circuit","level":3,"score":0.2773999869823456},{"id":"https://openalex.org/C148204187","wikidata":"https://www.wikidata.org/wiki/Q176541","display_name":"Optical computing","level":2,"score":0.26829999685287476}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2602.19031","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2602.19031","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":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.2602.19031","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2602.19031","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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"sustainable_development_goals":[{"id":"https://metadata.un.org/sdg/7","score":0.7484065890312195,"display_name":"Affordable and clean energy"}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"The":[0],"growing":[1],"computational":[2],"demands":[3],"of":[4],"artificial":[5],"intelligence":[6],"(AI)":[7],"are":[8],"challenging":[9],"conventional":[10],"electronics,":[11],"making":[12],"photonic":[13,20,35,229],"computing":[14],"a":[15,32,55,65,71,81,86,125,135,223],"promising":[16],"alternative.":[17],"However,":[18],"existing":[19],"architectures":[21],"face":[22],"fundamental":[23],"scalability":[24],"and":[25,52,90,138,156,200,208],"reliability":[26],"barriers.":[27],"This":[28],"paper":[29],"introduces":[30],"SKYLIGHT,":[31],"scalable":[33],"3D":[34,56,67],"in-memory":[36],"tensor":[37],"core":[38,131],"architecture":[39],"designed":[40],"for":[41,119,226],"real-time":[42,177],"AI":[43,230],"inference.":[44,116],"By":[45],"co-designing":[46],"its":[47,220],"topology,":[48,70],"wavelength":[49],"routing,":[50],"accumulation,":[51],"programming":[53],"in":[54,113,176],"stack,":[57],"SKYLIGHT":[58,99,130,214],"overcomes":[59],"key":[60],"limitations.":[61],"Its":[62],"innovations":[63],"include":[64],"low-loss":[66],"Si/SiN":[68],"crossbar":[69],"thermally":[72],"robust":[73],"non-micro-ring":[74],"resonator":[75],"(MRR)-based":[76],"wavelength-division":[77],"multiplexing":[78],"(WDM)":[79],"component,":[80],"hierarchical":[82],"signal":[83],"accumulation":[84],"using":[85],"multi-port":[87],"photodetector":[88],"(PD),":[89],"optically":[91],"programmed":[92],"non-volatile":[93],"phase-change":[94],"material":[95],"(PCM)":[96],"weights.":[97],"Importantly,":[98],"enables":[100],"in-situ":[101],"weight":[102],"updates":[103],"that":[104,124],"support":[105],"label-free,":[106],"layer-local":[107],"learning":[108,185],"(e.g.,":[109],"forward-forward":[110],"local":[111,189],"updates)":[112],"addition":[114],"to":[115,194],"Using":[117],"SimPhony":[118],"system-level":[120],"modeling,":[121],"we":[122],"show":[123],"single":[126,136],"144":[127],"x":[128,162],"256":[129],"is":[132],"feasible":[133],"within":[134],"reticle":[137],"delivers":[139],"342.1":[140],"TOPS":[141],"at":[142,148],"23.7":[143],"TOPS/W,":[144],"enabling":[145],"ResNet-50":[146],"inference":[147],"1212":[149],"FPS":[150],"with":[151],"27":[152],"mJ":[153],"per":[154],"image,":[155],"achieves":[157],"84.17":[158],"FPS/W":[159],"end-to-end":[160],"(1.61":[161],"higher":[163],"than":[164],"an":[165],"NVIDIA":[166],"RTX":[167],"PRO":[168],"6000":[169],"Blackwell":[170],"GPU)":[171],"under":[172],"the":[173],"same":[174],"workload":[175],"measurements.":[178],"System-level":[179],"evaluations":[180],"on":[181],"four":[182],"representative":[183],"machine":[184],"tasks,":[186],"including":[187],"unsupervised":[188],"self-learning,":[190],"demonstrate":[191],"SKYLIGHT's":[192],"robustness":[193],"realistic":[195],"hardware":[196],"non-idealities":[197],"(low-bit":[198],"quantization":[199],"signal-proportional":[201],"analog":[202],"noise":[203],"capturing":[204],"modulation,":[205],"PCM":[206],"programming,":[207],"readout":[209],"variations).":[210],"With":[211],"noise-aware":[212],"training,":[213],"maintains":[215],"high":[216],"task":[217],"accuracy,":[218],"validating":[219],"potential":[221],"as":[222],"comprehensive":[224],"solution":[225],"energy-efficient,":[227],"large-scale":[228],"accelerators.":[231]},"counts_by_year":[],"updated_date":"2026-07-28T07:46:37.118299","created_date":"2026-02-26T00:00:00"}