{"id":"https://openalex.org/W7021247103","doi":null,"title":"Norm-Based Generalization Bounds for Compositionally Sparse Neural Network","display_name":"Norm-Based Generalization Bounds for Compositionally Sparse Neural Network","publication_year":2023,"publication_date":"2023-02-14","ids":{"openalex":"https://openalex.org/W7021247103"},"language":"en","primary_location":{"id":"pmh:oai:dspace.mit.edu:1721.1/148230","doi":null,"is_oa":true,"landing_page_url":"https://hdl.handle.net/1721.1/148230","pdf_url":"https://dspace.mit.edu/bitstream/1721.1/148230/1/CBMM-Memo-139.pdf","source":{"id":"https://openalex.org/S4306400425","display_name":"DSpace@MIT (Massachusetts Institute of Technology)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I63966007","host_organization_name":"Massachusetts Institute of Technology","host_organization_lineage":["https://openalex.org/I63966007"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by-nc","license_id":"https://openalex.org/licenses/cc-by-nc","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Technical Report"},"volume":null,"type":"report","is_traditional":false,"indexed_in":[],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://dspace.mit.edu/bitstream/1721.1/148230/1/CBMM-Memo-139.pdf","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5050776111","display_name":"Tomer Galanti","orcid":null,"observed_orcids":[]},"institutions":[{"id":"https://openalex.org/I63966007","display_name":"Massachusetts Institute of Technology","ror":"https://ror.org/042nb2s44","country_code":"US","type":"education","lineage":["https://openalex.org/I63966007"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Galanti, Tomer","raw_affiliation_strings":[": Center for Brains, Minds, and Machines , MIT , Cambridge , MA , USA","Massachusetts Institute of Technol- ogy"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":": Center for Brains, Minds, and Machines , MIT , Cambridge , MA , USA","institution_ids":["https://openalex.org/I63966007"]},{"raw_affiliation_string":"Massachusetts Institute of Technol- ogy","institution_ids":["https://openalex.org/I63966007"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5102880001","display_name":"Mengjia Xu","orcid":"https://orcid.org/0000-0002-3627-8261","observed_orcids":["https://orcid.org/0000-0002-3627-8261"]},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Xu, Mengjia","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5132684966","display_name":"Liane Galanti","orcid":"https://orcid.org/0009-0005-3808-016X","observed_orcids":["https://orcid.org/0009-0005-3808-016X"]},"institutions":[{"id":"https://openalex.org/I16391192","display_name":"Tel Aviv University","ror":"https://ror.org/04mhzgx49","country_code":"IL","type":"education","lineage":["https://openalex.org/I16391192"]}],"countries":["IL"],"is_corresponding":false,"raw_author_name":"Galanti, Liane","raw_affiliation_strings":[": School of Computer Science , Tel Aviv University , Tel Aviv , Israel","Tel-Aviv University"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":": School of Computer Science , Tel Aviv University , Tel Aviv , Israel","institution_ids":["https://openalex.org/I16391192"]},{"raw_affiliation_string":"Tel-Aviv University","institution_ids":["https://openalex.org/I16391192"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5001833084","display_name":"Tomaso Poggio","orcid":"https://orcid.org/0000-0002-3944-0455","observed_orcids":["https://orcid.org/0000-0002-3944-0455"]},"institutions":[{"id":"https://openalex.org/I63966007","display_name":"Massachusetts Institute of Technology","ror":"https://ror.org/042nb2s44","country_code":"US","type":"education","lineage":["https://openalex.org/I63966007"]}],"countries":["US"],"is_corresponding":false,"raw_author_name":"Poggio, Tomaso","raw_affiliation_strings":[": Center for Brains, Minds, and Machines , MIT , Cambridge , MA , USA","Massachusetts Institute of Technol- ogy"],"raw_orcid":null,"affiliations":[{"raw_affiliation_string":": Center for Brains, Minds, and Machines , MIT , Cambridge , MA , USA","institution_ids":["https://openalex.org/I63966007"]},{"raw_affiliation_string":"Massachusetts Institute of Technol- ogy","institution_ids":["https://openalex.org/I63966007"]}]}],"institutions":[],"countries_distinct_count":2,"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":true,"primary_topic":{"id":"https://openalex.org/T11612","display_name":"Stochastic Gradient Optimization Techniques","score":0.8270999789237976,"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/T11612","display_name":"Stochastic Gradient Optimization Techniques","score":0.8270999789237976,"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/T10320","display_name":"Neural Networks and Applications","score":0.13930000364780426,"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/T12072","display_name":"Machine Learning and Algorithms","score":0.01269999984651804,"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/rademacher-complexity","display_name":"Rademacher complexity","score":0.8840000033378601},{"id":"https://openalex.org/keywords/generalization-bounds","display_name":"generalization bounds","score":0.9729999899864197},{"id":"https://openalex.org/keywords/sparse-neural-networks","display_name":"sparse neural networks","score":0.9079999923706055},{"id":"https://openalex.org/keywords/relu-neural-networks","display_name":"ReLU networks","score":0.7870000004768372},{"id":"https://openalex.org/keywords/convolutional-neural-networks","display_name":"convolutional neural networks","score":0.9959999918937683},{"id":"https://openalex.org/keywords/weight-sharing","display_name":"parameter sharing","score":0.765999972820282},{"id":"https://openalex.org/keywords/classification","display_name":"classification","score":0.5009999871253967}],"concepts":[{"id":"https://openalex.org/C177148314","wikidata":"https://www.wikidata.org/wiki/Q170084","display_name":"Generalization","level":2,"score":0.848800003528595},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.6075999736785889},{"id":"https://openalex.org/C81363708","wikidata":"https://www.wikidata.org/wiki/Q17084460","display_name":"Convolutional neural network","level":2,"score":0.5995000004768372},{"id":"https://openalex.org/C2780586882","wikidata":"https://www.wikidata.org/wiki/Q7520643","display_name":"Simple (philosophy)","level":2,"score":0.5544999837875366},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5335000157356262},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.5060999989509583},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4555000066757202},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.44179999828338623},{"id":"https://openalex.org/C14036430","wikidata":"https://www.wikidata.org/wiki/Q3736076","display_name":"Function (biology)","level":2,"score":0.4090000092983246},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.39399999380111694},{"id":"https://openalex.org/C2984842247","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep neural networks","level":3,"score":0.39320001006126404},{"id":"https://openalex.org/C38365724","wikidata":"https://www.wikidata.org/wiki/Q4677469","display_name":"Activation function","level":3,"score":0.3066999912261963},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.2865999937057495},{"id":"https://openalex.org/C148043351","wikidata":"https://www.wikidata.org/wiki/Q4456944","display_name":"Current (fluid)","level":2,"score":0.275299996137619},{"id":"https://openalex.org/C77637269","wikidata":"https://www.wikidata.org/wiki/Q7002051","display_name":"Neural coding","level":2,"score":0.2619999945163727},{"id":"https://openalex.org/C106131492","wikidata":"https://www.wikidata.org/wiki/Q3072260","display_name":"Filter (signal processing)","level":2,"score":0.26080000400543213},{"id":"https://openalex.org/C2776401178","wikidata":"https://www.wikidata.org/wiki/Q12050496","display_name":"Feature (linguistics)","level":2,"score":0.25609999895095825},{"id":"https://openalex.org/C77618280","wikidata":"https://www.wikidata.org/wiki/Q1155772","display_name":"Scheme (mathematics)","level":2,"score":0.2547000050544739},{"id":"https://openalex.org/C147710293","wikidata":"https://www.wikidata.org/wiki/Q849428","display_name":"Toeplitz matrix","level":2,"score":0.24699999392032623},{"id":"https://openalex.org/C77553402","wikidata":"https://www.wikidata.org/wiki/Q13222579","display_name":"Upper and lower bounds","level":2,"score":0.2434999942779541},{"id":"https://openalex.org/C12426560","wikidata":"https://www.wikidata.org/wiki/Q189569","display_name":"Basis (linear algebra)","level":2,"score":0.24009999632835388},{"id":"https://openalex.org/C118615104","wikidata":"https://www.wikidata.org/wiki/Q121416","display_name":"Discrete mathematics","level":1,"score":0.23430000245571136},{"id":"https://openalex.org/C45347329","wikidata":"https://www.wikidata.org/wiki/Q5166604","display_name":"Convolution (computer science)","level":3,"score":0.2337000072002411},{"id":"https://openalex.org/C179799912","wikidata":"https://www.wikidata.org/wiki/Q205084","display_name":"Computational complexity theory","level":2,"score":0.22769999504089355},{"id":"https://openalex.org/C2776502983","wikidata":"https://www.wikidata.org/wiki/Q690182","display_name":"Contrast (vision)","level":2,"score":0.2207999974489212},{"id":"https://openalex.org/C41608201","wikidata":"https://www.wikidata.org/wiki/Q980509","display_name":"Embedding","level":2,"score":0.21170000731945038},{"id":"https://openalex.org/C191795146","wikidata":"https://www.wikidata.org/wiki/Q3878446","display_name":"Norm (philosophy)","level":2,"score":0.19900000095367432},{"id":"https://openalex.org/C119322782","wikidata":"https://www.wikidata.org/wiki/Q2662236","display_name":"VC dimension","level":2,"score":0.1981000006198883},{"id":"https://openalex.org/C175202392","wikidata":"https://www.wikidata.org/wiki/Q2434543","display_name":"Time delay neural network","level":3,"score":0.1873999983072281},{"id":"https://openalex.org/C136197465","wikidata":"https://www.wikidata.org/wiki/Q1729295","display_name":"Variety (cybernetics)","level":2,"score":0.1873999983072281},{"id":"https://openalex.org/C114614502","wikidata":"https://www.wikidata.org/wiki/Q76592","display_name":"Combinatorics","level":1,"score":0.18559999763965607},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.18240000307559967},{"id":"https://openalex.org/C34388435","wikidata":"https://www.wikidata.org/wiki/Q2267362","display_name":"Bounded function","level":2,"score":0.18209999799728394},{"id":"https://openalex.org/C5917680","wikidata":"https://www.wikidata.org/wiki/Q2621825","display_name":"Basis function","level":2,"score":0.18050000071525574},{"id":"https://openalex.org/C12713177","wikidata":"https://www.wikidata.org/wiki/Q1900281","display_name":"Perspective (graphical)","level":2,"score":0.1784999966621399},{"id":"https://openalex.org/C45374587","wikidata":"https://www.wikidata.org/wiki/Q12525525","display_name":"Computation","level":2,"score":0.1762000024318695},{"id":"https://openalex.org/C2986949344","wikidata":"https://www.wikidata.org/wiki/Q9404","display_name":"Neural system","level":2,"score":0.17509999871253967},{"id":"https://openalex.org/C56372850","wikidata":"https://www.wikidata.org/wiki/Q1050404","display_name":"Sparse matrix","level":3,"score":0.17299999296665192},{"id":"https://openalex.org/C134466208","wikidata":"https://www.wikidata.org/wiki/Q1520657","display_name":"Weight function","level":2,"score":0.17260000109672546},{"id":"https://openalex.org/C204323151","wikidata":"https://www.wikidata.org/wiki/Q905424","display_name":"Range (aeronautics)","level":2,"score":0.17260000109672546}],"mesh":[],"locations_count":1,"locations":[{"id":"pmh:oai:dspace.mit.edu:1721.1/148230","doi":null,"is_oa":true,"landing_page_url":"https://hdl.handle.net/1721.1/148230","pdf_url":"https://dspace.mit.edu/bitstream/1721.1/148230/1/CBMM-Memo-139.pdf","source":{"id":"https://openalex.org/S4306400425","display_name":"DSpace@MIT (Massachusetts Institute of Technology)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I63966007","host_organization_name":"Massachusetts Institute of Technology","host_organization_lineage":["https://openalex.org/I63966007"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by-nc","license_id":"https://openalex.org/licenses/cc-by-nc","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Technical Report"}],"best_oa_location":{"id":"pmh:oai:dspace.mit.edu:1721.1/148230","doi":null,"is_oa":true,"landing_page_url":"https://hdl.handle.net/1721.1/148230","pdf_url":"https://dspace.mit.edu/bitstream/1721.1/148230/1/CBMM-Memo-139.pdf","source":{"id":"https://openalex.org/S4306400425","display_name":"DSpace@MIT (Massachusetts Institute of Technology)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"listed_in":[],"host_organization":"https://openalex.org/I63966007","host_organization_name":"Massachusetts Institute of Technology","host_organization_lineage":["https://openalex.org/I63966007"],"host_organization_lineage_names":[],"type":"repository"},"license":"cc-by-nc","license_id":"https://openalex.org/licenses/cc-by-nc","version":"submittedVersion","is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Technical Report"},"sustainable_development_goals":[],"study_designs":[],"awards":[{"id":"https://openalex.org/G2830697258","display_name":"Collaborative Proposal:   Object and Action Recognition in Time Sequences of Images:   Computational Neuroscience and Neurophysiology","funder_award_id":"0827427","funder_id":"https://openalex.org/F4320306076","funder_display_name":"National Science Foundation"},{"id":"https://openalex.org/G5998761286","display_name":"Computational Models and Physiological Studies of Feedback in Visual Object Recognition Tasks","funder_award_id":"0640097","funder_id":"https://openalex.org/F4320306076","funder_display_name":"National Science Foundation"},{"id":"https://openalex.org/G7201050375","display_name":"A Center for Brains, Minds and Machines: the Science and the Technology of Intelligence","funder_award_id":"1231216","funder_id":"https://openalex.org/F4320306076","funder_display_name":"National Science Foundation"},{"id":"https://openalex.org/G7809643752","display_name":null,"funder_award_id":"CCF-1231216","funder_id":"https://openalex.org/F4320324608","funder_display_name":"Companhia Brasileira de Metalurgia e Minera\u00e7\u00e3o"}],"funders":[{"id":"https://openalex.org/F4320306076","display_name":"National Science Foundation","ror":"https://ror.org/021nxhr62"},{"id":"https://openalex.org/F4320324608","display_name":"Companhia Brasileira de Metalurgia e Minera\u00e7\u00e3o","ror":"https://ror.org/03k8d9631"}],"has_content":{"pdf":true,"grobid_xml":true},"content_urls":{"pdf":"https://content.openalex.org/works/W7021247103.pdf","grobid_xml":"https://content.openalex.org/works/W7021247103.grobid-xml"},"referenced_works_count":15,"referenced_works":["https://openalex.org/W2112796928","https://openalex.org/W2194775991","https://openalex.org/W2257979135","https://openalex.org/W2518108298","https://openalex.org/W2766099245","https://openalex.org/W2777138330","https://openalex.org/W2911867426","https://openalex.org/W2946840143","https://openalex.org/W2971069898","https://openalex.org/W2995896684","https://openalex.org/W3036167779","https://openalex.org/W3094502228","https://openalex.org/W3173512790","https://openalex.org/W4245577611","https://openalex.org/W6966494678"],"related_works":["https://openalex.org/W4318751297","https://openalex.org/W4383173863","https://openalex.org/W7133215319","https://openalex.org/W7102238376","https://openalex.org/W4366565368","https://openalex.org/W7162453115","https://openalex.org/W2944820194","https://openalex.org/W3173512790","https://openalex.org/W4313483655","https://openalex.org/W4300546810"],"abstract_inverted_index":{"In":[0],"this":[1],"paper,":[2],"we":[3,67],"investigate":[4],"the":[5,45,48,53,56,107,114],"Rademacher":[6],"complexity":[7],"of":[8,20,47,52,55,61,75,106,116],"deep":[9,117],"sparse":[10,28],"neural":[11,30,34,118],"networks,":[12,31],"where":[13],"each":[14],"neuron":[15],"receives":[16],"a":[17],"small":[18],"number":[19],"inputs.":[21],"We":[22],"prove":[23],"generalization":[24,82,92],"bounds":[25,37,71,83],"for":[26],"multilayered":[27],"ReLU":[29],"including":[32],"convolutional":[33,49],"networks.":[35,119],"These":[36],"differ":[38],"from":[39],"previous":[40],"ones,":[41],"as":[42],"they":[43,86],"consider":[44],"norms":[46,54],"filters":[50],"instead":[51],"associated":[57],"Toeplitz":[58],"matrices,":[59],"independently":[60],"weight":[62],"sharing":[63],"between":[64],"neurons.":[65],"As":[66],"show":[68],"theoretically,":[69],"these":[70,100],"may":[72],"be":[73],"orders":[74],"magnitude":[76],"better":[77],"than":[78],"standard":[79],"norm-":[80],"based":[81],"and":[84],"empirically,":[85],"are":[87],"almost":[88],"non-vacuous":[89],"in":[90,93],"estimating":[91],"various":[94],"simple":[95],"classification":[96],"problems.":[97],"Taken":[98],"together,":[99],"results":[101],"suggest":[102],"that":[103],"compositional":[104],"sparsity":[105],"underlying":[108],"target":[109],"function":[110],"is":[111],"critical":[112],"to":[113],"success":[115]},"counts_by_year":[],"updated_date":"2026-11-07T11:25:07.637180","created_date":"2025-10-10T00:00:00"}