{"id":"https://openalex.org/W7168605044","doi":"https://doi.org/10.1145/3799902.3811231","title":"Taming optimization variance in compact neural shading networks","display_name":"Taming optimization variance in compact neural shading networks","publication_year":2026,"publication_date":"2026-07-16","ids":{"openalex":"https://openalex.org/W7168605044","doi":"https://doi.org/10.1145/3799902.3811231"},"language":null,"primary_location":{"id":"doi:10.1145/3799902.3811231","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3799902.3811231","pdf_url":null,"source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the Special Interest Group on Computer Graphics and Interactive Techniques Conference Conference Papers","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":true,"oa_status":"gold","oa_url":"https://doi.org/10.1145/3799902.3811231","any_repository_has_fulltext":null},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5079958610","display_name":"Benedikt Bitterli","orcid":"https://orcid.org/0000-0002-8799-7119"},"institutions":[{"id":"https://openalex.org/I4210127875","display_name":"Nvidia (United States)","ror":"https://ror.org/03jdj4y14","country_code":"US","type":"company","lineage":["https://openalex.org/I4210127875"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Benedikt Bitterli","raw_affiliation_strings":["NVIDIA, Redmond, USA"],"raw_orcid":"https://orcid.org/0000-0002-8799-7119","affiliations":[{"raw_affiliation_string":"NVIDIA, Redmond, USA","institution_ids":["https://openalex.org/I4210127875"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5026516592","display_name":"Petrik Clarberg","orcid":"https://orcid.org/0009-0004-9870-6307"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Petrik Clarberg","raw_affiliation_strings":["NVIDIA, Lund, Sweden"],"raw_orcid":"https://orcid.org/0009-0004-9870-6307","affiliations":[{"raw_affiliation_string":"NVIDIA, Lund, Sweden","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5141021267","display_name":"Chris Cummings","orcid":"https://orcid.org/0009-0006-9281-8644"},"institutions":[{"id":"https://openalex.org/I1304085615","display_name":"Nvidia (United Kingdom)","ror":"https://ror.org/02kr42612","country_code":"GB","type":"company","lineage":["https://openalex.org/I1304085615","https://openalex.org/I4210127875"]}],"countries":["GB"],"is_corresponding":false,"raw_author_name":"Chris Cummings","raw_affiliation_strings":["NVIDIA, Guildford, United Kingdom"],"raw_orcid":"https://orcid.org/0009-0006-9281-8644","affiliations":[{"raw_affiliation_string":"NVIDIA, Guildford, United Kingdom","institution_ids":["https://openalex.org/I1304085615"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5087647732","display_name":"Aaron Lefohn","orcid":"https://orcid.org/0009-0002-6526-0922"},"institutions":[{"id":"https://openalex.org/I4210127875","display_name":"Nvidia (United States)","ror":"https://ror.org/03jdj4y14","country_code":"US","type":"company","lineage":["https://openalex.org/I4210127875"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Aaron Lefohn","raw_affiliation_strings":["NVIDIA, Redmond, USA"],"raw_orcid":"https://orcid.org/0009-0002-6526-0922","affiliations":[{"raw_affiliation_string":"NVIDIA, Redmond, USA","institution_ids":["https://openalex.org/I4210127875"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5031672248","display_name":"Steve Marschner","orcid":"https://orcid.org/0000-0001-9810-0306"},"institutions":[{"id":"https://openalex.org/I205783295","display_name":"Cornell University","ror":"https://ror.org/05bnh6r87","country_code":"US","type":"education","lineage":["https://openalex.org/I205783295"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Steve Marschner","raw_affiliation_strings":["Cornell University, Ithaca, USA and NVIDIA, Ithaca, USA"],"raw_orcid":"https://orcid.org/0000-0001-9810-0306","affiliations":[{"raw_affiliation_string":"Cornell University, Ithaca, USA and NVIDIA, Ithaca, USA","institution_ids":["https://openalex.org/I205783295"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5050969104","display_name":"Jan Nov\u00e1k","orcid":"https://orcid.org/0000-0002-8320-9584"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Jan Nov\u00e1k","raw_affiliation_strings":["NVIDIA, Prague, Czech Republic"],"raw_orcid":"https://orcid.org/0000-0002-8320-9584","affiliations":[{"raw_affiliation_string":"NVIDIA, Prague, Czech Republic","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5037043205","display_name":"Fabrice Rousselle","orcid":"https://orcid.org/0009-0003-2978-2130"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Fabrice Rousselle","raw_affiliation_strings":["NVIDIA, Zurich, Switzerland"],"raw_orcid":"https://orcid.org/0009-0003-2978-2130","affiliations":[{"raw_affiliation_string":"NVIDIA, Zurich, Switzerland","institution_ids":[]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5007428449","display_name":"Andrea Weidlich","orcid":"https://orcid.org/0000-0002-4146-187X"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Andrea Weidlich","raw_affiliation_strings":["NVIDIA, Montreal, Canada"],"raw_orcid":"https://orcid.org/0000-0002-4146-187X","affiliations":[{"raw_affiliation_string":"NVIDIA, Montreal, Canada","institution_ids":[]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5010255059","display_name":"Tizian Zeltner","orcid":"https://orcid.org/0000-0002-6434-465X"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Tizian Zeltner","raw_affiliation_strings":["NVIDIA, Zurich, Switzerland"],"raw_orcid":"https://orcid.org/0000-0002-6434-465X","affiliations":[{"raw_affiliation_string":"NVIDIA, Zurich, Switzerland","institution_ids":[]}]}],"institutions":[],"countries_distinct_count":2,"institutions_distinct_count":3,"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":"1","last_page":"9"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":null,"topics":[],"keywords":[{"id":"https://openalex.org/keywords/artificial-neural-network","display_name":"Artificial neural network","score":0.7232999801635742},{"id":"https://openalex.org/keywords/context","display_name":"Context (archaeology)","score":0.6175000071525574},{"id":"https://openalex.org/keywords/range","display_name":"Range (aeronautics)","score":0.5407999753952026},{"id":"https://openalex.org/keywords/training","display_name":"Training (meteorology)","score":0.5116999745368958},{"id":"https://openalex.org/keywords/optimization-problem","display_name":"Optimization problem","score":0.40149998664855957},{"id":"https://openalex.org/keywords/task","display_name":"Task (project management)","score":0.3880999982357025},{"id":"https://openalex.org/keywords/variance","display_name":"Variance (accounting)","score":0.37389999628067017},{"id":"https://openalex.org/keywords/quality","display_name":"Quality (philosophy)","score":0.3416000008583069}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7335000038146973},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.7232999801635742},{"id":"https://openalex.org/C2779343474","wikidata":"https://www.wikidata.org/wiki/Q3109175","display_name":"Context (archaeology)","level":2,"score":0.6175000071525574},{"id":"https://openalex.org/C204323151","wikidata":"https://www.wikidata.org/wiki/Q905424","display_name":"Range (aeronautics)","level":2,"score":0.5407999753952026},{"id":"https://openalex.org/C2777211547","wikidata":"https://www.wikidata.org/wiki/Q17141490","display_name":"Training (meteorology)","level":2,"score":0.5116999745368958},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.48080000281333923},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.41029998660087585},{"id":"https://openalex.org/C137836250","wikidata":"https://www.wikidata.org/wiki/Q984063","display_name":"Optimization problem","level":2,"score":0.40149998664855957},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.3880999982357025},{"id":"https://openalex.org/C196083921","wikidata":"https://www.wikidata.org/wiki/Q7915758","display_name":"Variance (accounting)","level":2,"score":0.37389999628067017},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.3504999876022339},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.34290000796318054},{"id":"https://openalex.org/C2779530757","wikidata":"https://www.wikidata.org/wiki/Q1207505","display_name":"Quality (philosophy)","level":2,"score":0.3416000008583069},{"id":"https://openalex.org/C2777303404","wikidata":"https://www.wikidata.org/wiki/Q759757","display_name":"Convergence (economics)","level":2,"score":0.32420000433921814},{"id":"https://openalex.org/C2984842247","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep neural networks","level":3,"score":0.31779998540878296},{"id":"https://openalex.org/C2987595161","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Optimization algorithm","level":2,"score":0.3091000020503998},{"id":"https://openalex.org/C51632099","wikidata":"https://www.wikidata.org/wiki/Q3985153","display_name":"Training set","level":2,"score":0.30469998717308044},{"id":"https://openalex.org/C147764199","wikidata":"https://www.wikidata.org/wiki/Q6865248","display_name":"Minification","level":2,"score":0.2989000082015991},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.2669000029563904},{"id":"https://openalex.org/C155032097","wikidata":"https://www.wikidata.org/wiki/Q798503","display_name":"Backpropagation","level":3,"score":0.2619999945163727},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.2612999975681305},{"id":"https://openalex.org/C14036430","wikidata":"https://www.wikidata.org/wiki/Q3736076","display_name":"Function (biology)","level":2,"score":0.2531999945640564}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.1145/3799902.3811231","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3799902.3811231","pdf_url":null,"source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the Special Interest Group on Computer Graphics and Interactive Techniques Conference Conference Papers","raw_type":"proceedings-article"}],"best_oa_location":{"id":"doi:10.1145/3799902.3811231","is_oa":true,"landing_page_url":"https://doi.org/10.1145/3799902.3811231","pdf_url":null,"source":null,"license":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the Special Interest Group on Computer Graphics and Interactive Techniques Conference Conference Papers","raw_type":"proceedings-article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":18,"referenced_works":["https://openalex.org/W1494645067","https://openalex.org/W2792287754","https://openalex.org/W2922301133","https://openalex.org/W2942074357","https://openalex.org/W2963527567","https://openalex.org/W2983368141","https://openalex.org/W2983597143","https://openalex.org/W2986102549","https://openalex.org/W3024211078","https://openalex.org/W3105515454","https://openalex.org/W3176006498","https://openalex.org/W3180432685","https://openalex.org/W3191375248","https://openalex.org/W3216062517","https://openalex.org/W4285981718","https://openalex.org/W4394980032","https://openalex.org/W4396926989","https://openalex.org/W7115013331"],"related_works":[],"abstract_inverted_index":{"We":[0,63,103],"present":[1],"a":[2,90,99,121],"training":[3,47,70,143,163],"algorithm":[4,108],"to":[5,40,72,129,140,158],"mitigate":[6],"optimization":[7,34,52],"instabilities":[8],"in":[9,16,109],"small":[10,125],"neural":[11,18,117,126],"networks,":[12],"like":[13],"those":[14],"used":[15],"real-time":[17],"shading":[19],"applications.":[20],"While":[21],"large,":[22],"overparameterized":[23],"models":[24,38,127],"exhibit":[25],"predictable":[26],"convergence,":[27],"smaller":[28],"architectures":[29],"often":[30],"suffer":[31],"from":[32],"high":[33,131],"variance:":[35],"differently":[36],"initialized":[37],"converge":[39],"disparate":[41],"local":[42],"minima.":[43],"To":[44],"reduce":[45],"these":[46],"instabilities,":[48],"we":[49,145],"introduce":[50],"an":[51,56],"approach":[53],"that":[54],"utilizes":[55],"ensemble":[57],"of":[58,112,134,150],"network":[59],"instances":[60,66],"during":[61],"training.":[62,80],"prune":[64],"underperforming":[65],"and":[67,105,155,162],"dynamically":[68],"resize":[69],"batches":[71],"maintain":[73],"wall-clock":[74],"timings":[75],"comparable":[76],"to\u2014or":[77],"faster":[78],"than\u2014single-instance":[79],"This":[81,119],"strategy":[82],"allows":[83],"efficiently":[84],"exploring":[85],"the":[86,107,110,130,135,148],"weight":[87],"space":[88],"yielding":[89],"well-performing":[91],"model":[92],"with":[93],"significantly":[94],"higher":[95],"likelihood":[96],"than":[97],"optimizing":[98],"single":[100],"instance":[101],"only.":[102],"develop":[104],"analyze":[106],"context":[111],"learning":[113],"reflectance":[114],"functions":[115],"for":[116,124],"shading.":[118],"is":[120],"challenging":[122],"task":[123],"due":[128],"dynamic":[132],"range":[133],"target":[136],"function.":[137],"In":[138],"addition":[139],"our":[141],"multi-instance":[142],"method,":[144],"also":[146],"revisit":[147],"choices":[149],"loss":[151],"functions,":[152,154],"activation":[153],"input":[156],"parameterization":[157],"further":[159],"improve":[160],"quality":[161],"robustness.":[164]},"counts_by_year":[],"updated_date":"2026-07-17T05:58:32.477365","created_date":"2026-07-17T00:00:00"}