{"id":"https://openalex.org/W7163138173","doi":"https://doi.org/10.48550/arxiv.2606.01544","title":"CRePE: Convolution-aware Relative Importance in Post-training Pruning with Efficient Search","display_name":"CRePE: Convolution-aware Relative Importance in Post-training Pruning with Efficient Search","publication_year":2026,"publication_date":"2026-06-01","ids":{"openalex":"https://openalex.org/W7163138173","doi":"https://doi.org/10.48550/arxiv.2606.01544"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2606.01544","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.01544","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","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.2606.01544","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5137630614","display_name":"Cheonjun Park","orcid":null},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Park, Cheonjun","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"corresponding_author_ids":["https://openalex.org/A5137630614"],"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/T10028","display_name":"Topic Modeling","score":0.4041000008583069,"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/T10028","display_name":"Topic Modeling","score":0.4041000008583069,"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/T10181","display_name":"Natural Language Processing Techniques","score":0.14139999449253082,"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/T11307","display_name":"Domain Adaptation and Few-Shot Learning","score":0.06199999898672104,"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/pruning","display_name":"Pruning","score":0.6482999920845032},{"id":"https://openalex.org/keywords/hyperparameter","display_name":"Hyperparameter","score":0.6205000281333923},{"id":"https://openalex.org/keywords/hill-climbing","display_name":"Hill climbing","score":0.5701000094413757},{"id":"https://openalex.org/keywords/context","display_name":"Context (archaeology)","score":0.48919999599456787},{"id":"https://openalex.org/keywords/perplexity","display_name":"Perplexity","score":0.4731000065803528},{"id":"https://openalex.org/keywords/column","display_name":"Column (typography)","score":0.4007999897003174}],"concepts":[{"id":"https://openalex.org/C108010975","wikidata":"https://www.wikidata.org/wiki/Q500094","display_name":"Pruning","level":2,"score":0.6482999920845032},{"id":"https://openalex.org/C8642999","wikidata":"https://www.wikidata.org/wiki/Q4171168","display_name":"Hyperparameter","level":2,"score":0.6205000281333923},{"id":"https://openalex.org/C135450995","wikidata":"https://www.wikidata.org/wiki/Q820272","display_name":"Hill climbing","level":2,"score":0.5701000094413757},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.5677000284194946},{"id":"https://openalex.org/C2779343474","wikidata":"https://www.wikidata.org/wiki/Q3109175","display_name":"Context (archaeology)","level":2,"score":0.48919999599456787},{"id":"https://openalex.org/C100279451","wikidata":"https://www.wikidata.org/wiki/Q372193","display_name":"Perplexity","level":3,"score":0.4731000065803528},{"id":"https://openalex.org/C2780551164","wikidata":"https://www.wikidata.org/wiki/Q2306599","display_name":"Column (typography)","level":3,"score":0.4007999897003174},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.39410001039505005},{"id":"https://openalex.org/C33923547","wikidata":"https://www.wikidata.org/wiki/Q395","display_name":"Mathematics","level":0,"score":0.38449999690055847},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3776000142097473},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.37540000677108765},{"id":"https://openalex.org/C177148314","wikidata":"https://www.wikidata.org/wiki/Q170084","display_name":"Generalization","level":2,"score":0.2921000123023987},{"id":"https://openalex.org/C126255220","wikidata":"https://www.wikidata.org/wiki/Q141495","display_name":"Mathematical optimization","level":1,"score":0.25619998574256897},{"id":"https://openalex.org/C127162648","wikidata":"https://www.wikidata.org/wiki/Q16858953","display_name":"Channel (broadcasting)","level":2,"score":0.25209999084472656},{"id":"https://openalex.org/C127705205","wikidata":"https://www.wikidata.org/wiki/Q5748245","display_name":"Heuristics","level":2,"score":0.25200000405311584}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2606.01544","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.01544","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"Preprint"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2606.01544","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2606.01544","pdf_url":null,"source":{"id":"https://openalex.org/S4306400194","display_name":"arXiv (Cornell University)","issn_l":null,"issn":null,"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":"cc-by","license_id":"https://openalex.org/licenses/cc-by","version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"Preprint"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Deploying":[0],"Large":[1],"Language":[2],"Models":[3],"(LLMs)":[4],"in":[5],"practice":[6],"incurs":[7],"substantial":[8],"memory":[9],"and":[10,41,56,62,77,93,110,134,180],"computational":[11],"costs.":[12],"Post-training":[13],"pruning":[14],"(PTP)":[15],"is":[16],"an":[17],"effective":[18],"approach":[19],"to":[20,60,139,156],"reducing":[21],"these":[22],"costs":[23],"by":[24,39,149],"removing":[25],"weights":[26],"without":[27],"additional":[28],"training.":[29],"Among":[30],"existing":[31,87,172],"methods,":[32],"RIA":[33,48],"introduces":[34],"relative":[35],"importance":[36],"scores":[37],"normalized":[38],"row":[40,61],"column":[42,63],"sums,":[43],"achieving":[44],"state-of-the-art":[45],"accuracy.":[46],"However,":[47,96],"considers":[49],"only":[50],"1D":[51],"cross-shaped":[52],"(row/column)":[53],"directional":[54],"information":[55],"assigns":[57],"equal":[58],"weight":[59],"contributions.":[64],"In":[65],"this":[66],"paper,":[67],"we":[68,120,163],"propose":[69,121],"\\textbf{CRePE},":[70],"which":[71,126],"incorporates":[72],"2D":[73],"local":[74],"neighborhood":[75],"context":[76],"adaptive":[78,99],"coefficients":[79,100],"into":[80],"Relative":[81],"Importance":[82],"scoring.":[83],"CRePE":[84,166],"consistently":[85],"outperforms":[86],"PTP":[88],"methods":[89],"across":[90],"diverse":[91],"models":[92],"sparsity":[94,178],"settings.":[95],"identifying":[97],"optimal":[98,145],"via":[101],"perplexity":[102],"(PPL)-based":[103],"hill":[104],"climbing":[105],"requires":[106],"numerous":[107],"PPL":[108,132],"evaluations":[109],"approximately":[111,140],"11":[112],"hours":[113],"of":[114],"search":[115,137],"time.":[116],"To":[117],"address":[118],"this,":[119],"\\textbf{PHO}":[122],"(Proxy-based":[123],"Hyperparameter":[124],"Optimization),":[125],"eliminates":[127],"the":[128,136,144],"need":[129],"for":[130],"repeated":[131],"measurements":[133],"reduces":[135],"time":[138],"20":[141],"minutes.":[142],"Furthermore,":[143],"hyperparameter":[146],"configuration":[147],"found":[148],"PHO":[150],"on":[151],"one":[152],"model":[153],"transfers":[154],"well":[155],"other":[157],"models,":[158],"demonstrating":[159],"strong":[160],"generalization.":[161],"Finally,":[162],"verify":[164],"that":[165],"can":[167],"be":[168],"orthogonally":[169],"combined":[170],"with":[171],"techniques":[173],"including":[174],"Channel":[175],"Permutation,":[176],"non-uniform":[177],"allocation,":[179],"re-pruning":[181],"methods.":[182]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-06-03T00:00:00"}