{"id":"https://openalex.org/W4401863784","doi":"https://doi.org/10.1145/3637528.3671776","title":"Learning from Emergence: A Study on Proactively Inhibiting the Monosemantic Neurons of Artificial Neural Networks","display_name":"Learning from Emergence: A Study on Proactively Inhibiting the Monosemantic Neurons of Artificial Neural Networks","publication_year":2024,"publication_date":"2024-08-24","ids":{"openalex":"https://openalex.org/W4401863784","doi":"https://doi.org/10.1145/3637528.3671776"},"language":"en","primary_location":{"id":"doi:10.1145/3637528.3671776","is_oa":false,"landing_page_url":"https://doi.org/10.1145/3637528.3671776","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining","raw_type":"proceedings-article"},"type":"conference-paper","indexed_in":["crossref"],"open_access":{"is_oa":false,"oa_status":"closed","oa_url":null,"any_repository_has_fulltext":false},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5004999121","display_name":"J Wang","orcid":"https://orcid.org/0000-0001-6473-8221"},"institutions":[{"id":"https://openalex.org/I200769079","display_name":"Hong Kong University of Science and Technology","ror":"https://ror.org/00q4vv597","country_code":"HK","type":"education","lineage":["https://openalex.org/I200769079"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Jiachuan Wang","raw_affiliation_strings":["The Hong Kong University of Science and Technology, Hong Kong SAR, China"],"raw_orcid":"https://orcid.org/0000-0001-6473-8221","affiliations":[{"raw_affiliation_string":"The Hong Kong University of Science and Technology, Hong Kong SAR, China","institution_ids":["https://openalex.org/I200769079"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5006260400","display_name":"Shimin Di","orcid":"https://orcid.org/0000-0002-7394-0082"},"institutions":[{"id":"https://openalex.org/I200769079","display_name":"Hong Kong University of Science and Technology","ror":"https://ror.org/00q4vv597","country_code":"HK","type":"education","lineage":["https://openalex.org/I200769079"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Shimin Di","raw_affiliation_strings":["The Hong Kong University of Science and Technology, Hong Kong SAR, China"],"raw_orcid":"https://orcid.org/0000-0002-7394-0082","affiliations":[{"raw_affiliation_string":"The Hong Kong University of Science and Technology, Hong Kong SAR, China","institution_ids":["https://openalex.org/I200769079"]}]},{"author_position":"middle","author":{"id":"https://openalex.org/A5100333516","display_name":"Lei Chen","orcid":"https://orcid.org/0000-0002-8257-5806"},"institutions":[{"id":"https://openalex.org/I200769079","display_name":"Hong Kong University of Science and Technology","ror":"https://ror.org/00q4vv597","country_code":"HK","type":"education","lineage":["https://openalex.org/I200769079"]},{"id":"https://openalex.org/I889458895","display_name":"University of Hong Kong","ror":"https://ror.org/02zhqgq86","country_code":"HK","type":"education","lineage":["https://openalex.org/I889458895"]}],"countries":["HK"],"is_corresponding":false,"raw_author_name":"Lei Chen","raw_affiliation_strings":["The Hong Kong University of Science and Technology, (Guangzhou), Guangzhou, China"],"raw_orcid":"https://orcid.org/0000-0002-8257-5806","affiliations":[{"raw_affiliation_string":"The Hong Kong University of Science and Technology, (Guangzhou), Guangzhou, China","institution_ids":["https://openalex.org/I200769079","https://openalex.org/I889458895"]}]},{"author_position":"last","author":{"id":"https://openalex.org/A5047314253","display_name":"Charles Wang Wai Ng","orcid":"https://orcid.org/0000-0001-6693-3151"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Charles Wang Wai Ng","raw_affiliation_strings":["The Hong Kong University of Science and Technology (Guangzhou), Guangzhou, China"],"raw_orcid":"https://orcid.org/0000-0001-6693-3151","affiliations":[{"raw_affiliation_string":"The Hong Kong University of Science and Technology (Guangzhou), Guangzhou, China","institution_ids":[]}]}],"institutions":[],"countries_distinct_count":1,"institutions_distinct_count":2,"corresponding_author_ids":[],"corresponding_institution_ids":[],"apc_list":null,"apc_paid":null,"fwci":null,"has_fulltext":false,"cited_by_count":1,"citation_normalized_percentile":null,"cited_by_percentile_year":null,"biblio":{"volume":null,"issue":null,"first_page":"3092","last_page":"3103"},"is_retracted":false,"is_paratext":false,"is_xpac":false,"primary_topic":{"id":"https://openalex.org/T10320","display_name":"Neural Networks and Applications","score":0.9883999824523926,"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/T10320","display_name":"Neural Networks and Applications","score":0.9883999824523926,"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/T10502","display_name":"Advanced Memory and Neural Computing","score":0.9861000180244446,"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/T11975","display_name":"Evolutionary Algorithms and Applications","score":0.961899995803833,"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/artificial-neural-network","display_name":"Artificial neural network","score":0.6842190623283386},{"id":"https://openalex.org/keywords/computer-science","display_name":"Computer science","score":0.6331086158752441},{"id":"https://openalex.org/keywords/artificial-intelligence","display_name":"Artificial intelligence","score":0.45932716131210327}],"concepts":[{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.6842190623283386},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6331086158752441},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.45932716131210327}],"mesh":[],"locations_count":2,"locations":[{"id":"doi:10.1145/3637528.3671776","is_oa":false,"landing_page_url":"https://doi.org/10.1145/3637528.3671776","pdf_url":null,"source":null,"license":null,"license_id":null,"version":"publishedVersion","is_accepted":true,"is_published":true,"raw_source_name":"Proceedings of the 30th ACM SIGKDD Conference on Knowledge Discovery and Data Mining","raw_type":"proceedings-article"},{"id":"pmh:oai:repository.hkust.edu.hk:1783.1-143576","is_oa":false,"landing_page_url":"http://repository.hkust.edu.hk/ir/Record/1783.1-143576","pdf_url":null,"source":{"id":"https://openalex.org/S4306401796","display_name":"Rare & Special e-Zone (The Hong Kong University of Science and Technology)","issn_l":null,"issn":null,"is_oa":false,"is_in_doaj":false,"is_core":false,"host_organization":"https://openalex.org/I200769079","host_organization_name":"Hong Kong University of Science and Technology","host_organization_lineage":["https://openalex.org/I200769079"],"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":"Conference paper"}],"best_oa_location":null,"sustainable_development_goals":[{"id":"https://metadata.un.org/sdg/10","score":0.4000000059604645,"display_name":"Reduced inequalities"}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":13,"referenced_works":["https://openalex.org/W2002011878","https://openalex.org/W2147800946","https://openalex.org/W2194775991","https://openalex.org/W2601622117","https://openalex.org/W3010694149","https://openalex.org/W3082665562","https://openalex.org/W3095319910","https://openalex.org/W3134095442","https://openalex.org/W3138516171","https://openalex.org/W4232900388","https://openalex.org/W4239072543","https://openalex.org/W4300435436","https://openalex.org/W4302400283"],"related_works":["https://openalex.org/W4391375266","https://openalex.org/W2748952813","https://openalex.org/W2390279801","https://openalex.org/W2358668433","https://openalex.org/W4396701345","https://openalex.org/W2376932109","https://openalex.org/W2001405890","https://openalex.org/W4396696052","https://openalex.org/W2382290278","https://openalex.org/W4395014643"],"abstract_inverted_index":{"Recently,":[0],"emergence":[1],"has":[2],"received":[3],"widespread":[4],"attention":[5],"from":[6,18],"the":[7,12,19,28,31,35,60,118,123,143,193],"research":[8],"community":[9],"along":[10],"with":[11,46,122],"success":[13],"of":[14,33,37,120,125,145,169,195],"large-scale":[15],"models.":[16,64],"Different":[17],"literature,":[20],"we":[21,69,110],"hypothesize":[22],"a":[23,86,113,132,167],"key":[24],"factor":[25],"that":[26,40,156],"promotes":[27],"performance":[29,61,160],"during":[30],"increase":[32],"scale:":[34],"reduction":[36],"monosemantic":[38,76,100,138],"neurons":[39,50,77,101,121,139,147],"can":[41],"only":[42],"form":[43],"one-to-one":[44],"correlations":[45],"specific":[47],"features.":[48],"Monosemantic":[49],"tend":[51],"to":[52,74,116,136],"be":[53],"sparser":[54],"and":[55,78,97,140,172,181,190],"have":[56],"negative":[57],"impacts":[58],"on":[59,166],"in":[62,106,148,175],"large":[63],"Inspired":[65],"by":[66],"this":[67,83],"insight,":[68],"propose":[70,112],"an":[71],"intuitive":[72],"idea":[73],"identify":[75],"inhibit":[79],"them.":[80],"However,":[81],"achieving":[82],"goal":[84],"is":[85,91],"non-trivial":[87],"task":[88],"as":[89],"there":[90],"no":[92],"unified":[93],"quantitative":[94],"evaluation":[95],"metric":[96,115],"simply":[98],"banning":[99],"does":[102],"not":[103],"promote":[104,142],"polysemanticity":[105],"neural":[107,150,170],"networks.":[108,151],"Therefore,":[109],"first":[111],"new":[114],"measure":[117],"monosemanticity":[119,157],"guarantee":[124],"efficiency":[126],"for":[127],"online":[128],"computation,":[129],"then":[130],"introduce":[131],"theoretically":[133],"supported":[134],"method":[135],"suppress":[137],"proactively":[141],"ratios":[144],"polysemantic":[146],"training":[149],"We":[152],"validate":[153,187],"our":[154,188],"conjecture":[155],"brings":[158],"about":[159],"change":[161],"at":[162],"different":[163,176],"model":[164],"scales":[165],"variety":[168],"networks":[171],"benchmark":[173],"datasets":[174],"areas,":[177],"including":[178],"language,":[179],"image,":[180],"physics":[182],"simulation":[183],"tasks.":[184],"Further":[185],"experiments":[186],"analysis":[189],"theory":[191],"regarding":[192],"inhibition":[194],"monosemanticity.":[196]},"counts_by_year":[{"year":2025,"cited_by_count":1}],"updated_date":"2026-07-14T23:27:15.235271","created_date":"2025-10-10T00:00:00"}