{"id":"https://openalex.org/W7147635758","doi":"https://doi.org/10.48550/arxiv.2603.26756","title":"GradAttn: Replacing Fixed Residual Connections with Task-Modulated Attention Pathways","display_name":"GradAttn: Replacing Fixed Residual Connections with Task-Modulated Attention Pathways","publication_year":2026,"publication_date":"2026-03-23","ids":{"openalex":"https://openalex.org/W7147635758","doi":"https://doi.org/10.48550/arxiv.2603.26756"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2603.26756","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.26756","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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"article"},"type":"preprint","indexed_in":["datacite"],"open_access":{"is_oa":true,"oa_status":"green","oa_url":"https://doi.org/10.48550/arxiv.2603.26756","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5075023182","display_name":"Sheela Ghoshal","orcid":null},"institutions":[],"countries":[],"is_corresponding":true,"raw_author_name":"Ghoshal, Soudeep","raw_affiliation_strings":[],"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5018753704","display_name":"Himanshu Buckchash","orcid":"https://orcid.org/0000-0003-3679-3498"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Buckchash, Himanshu","raw_affiliation_strings":[],"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":2,"corresponding_author_ids":["https://openalex.org/A5075023182"],"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/T10775","display_name":"Generative Adversarial Networks and Image Synthesis","score":0.43479999899864197,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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/T10775","display_name":"Generative Adversarial Networks and Image Synthesis","score":0.43479999899864197,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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":0.11540000140666962,"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/T11448","display_name":"Face recognition and analysis","score":0.06949999928474426,"subfield":{"id":"https://openalex.org/subfields/1707","display_name":"Computer Vision and Pattern Recognition"},"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/encoding","display_name":"Encoding (memory)","score":0.7080000042915344},{"id":"https://openalex.org/keywords/residual","display_name":"Residual","score":0.699400007724762},{"id":"https://openalex.org/keywords/representation","display_name":"Representation (politics)","score":0.6467999815940857},{"id":"https://openalex.org/keywords/feature","display_name":"Feature (linguistics)","score":0.5480999946594238},{"id":"https://openalex.org/keywords/deep-learning","display_name":"Deep learning","score":0.5224000215530396},{"id":"https://openalex.org/keywords/stability","display_name":"Stability (learning theory)","score":0.49470001459121704},{"id":"https://openalex.org/keywords/pattern-recognition","display_name":"Pattern recognition (psychology)","score":0.48559999465942383},{"id":"https://openalex.org/keywords/task","display_name":"Task (project management)","score":0.4781999886035919}],"concepts":[{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7099000215530396},{"id":"https://openalex.org/C125411270","wikidata":"https://www.wikidata.org/wiki/Q18653","display_name":"Encoding (memory)","level":2,"score":0.7080000042915344},{"id":"https://openalex.org/C155512373","wikidata":"https://www.wikidata.org/wiki/Q287450","display_name":"Residual","level":2,"score":0.699400007724762},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.6917999982833862},{"id":"https://openalex.org/C2776359362","wikidata":"https://www.wikidata.org/wiki/Q2145286","display_name":"Representation (politics)","level":3,"score":0.6467999815940857},{"id":"https://openalex.org/C2776401178","wikidata":"https://www.wikidata.org/wiki/Q12050496","display_name":"Feature (linguistics)","level":2,"score":0.5480999946594238},{"id":"https://openalex.org/C108583219","wikidata":"https://www.wikidata.org/wiki/Q197536","display_name":"Deep learning","level":2,"score":0.5224000215530396},{"id":"https://openalex.org/C112972136","wikidata":"https://www.wikidata.org/wiki/Q7595718","display_name":"Stability (learning theory)","level":2,"score":0.49470001459121704},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.48559999465942383},{"id":"https://openalex.org/C2780451532","wikidata":"https://www.wikidata.org/wiki/Q759676","display_name":"Task (project management)","level":2,"score":0.4781999886035919},{"id":"https://openalex.org/C188198153","wikidata":"https://www.wikidata.org/wiki/Q1613840","display_name":"Limiting","level":2,"score":0.4221999943256378},{"id":"https://openalex.org/C50644808","wikidata":"https://www.wikidata.org/wiki/Q192776","display_name":"Artificial neural network","level":2,"score":0.3991999924182892},{"id":"https://openalex.org/C59404180","wikidata":"https://www.wikidata.org/wiki/Q17013334","display_name":"Feature learning","level":2,"score":0.3763999938964844},{"id":"https://openalex.org/C116409475","wikidata":"https://www.wikidata.org/wiki/Q1385056","display_name":"External Data Representation","level":2,"score":0.35830000042915344},{"id":"https://openalex.org/C2779843651","wikidata":"https://www.wikidata.org/wiki/Q7390335","display_name":"SIGNAL (programming language)","level":2,"score":0.3580999970436096},{"id":"https://openalex.org/C52622490","wikidata":"https://www.wikidata.org/wiki/Q1026626","display_name":"Feature extraction","level":2,"score":0.3285999894142151},{"id":"https://openalex.org/C193415008","wikidata":"https://www.wikidata.org/wiki/Q639681","display_name":"Network architecture","level":2,"score":0.3248000144958496},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.3163999915122986},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.3138999938964844},{"id":"https://openalex.org/C77618280","wikidata":"https://www.wikidata.org/wiki/Q1155772","display_name":"Scheme (mathematics)","level":2,"score":0.30820000171661377},{"id":"https://openalex.org/C2944601119","wikidata":"https://www.wikidata.org/wiki/Q43744058","display_name":"Residual neural network","level":3,"score":0.30070000886917114},{"id":"https://openalex.org/C104267543","wikidata":"https://www.wikidata.org/wiki/Q208163","display_name":"Signal processing","level":3,"score":0.27469998598098755},{"id":"https://openalex.org/C167879884","wikidata":"https://www.wikidata.org/wiki/Q727568","display_name":"Balanced flow","level":2,"score":0.2718999981880188},{"id":"https://openalex.org/C175154964","wikidata":"https://www.wikidata.org/wiki/Q380077","display_name":"Task analysis","level":3,"score":0.26919999718666077}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2603.26756","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.26756","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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":null,"raw_source_name":null,"raw_type":"article"}],"best_oa_location":{"id":"doi:10.48550/arxiv.2603.26756","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2603.26756","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":null,"license_id":null,"version":null,"is_accepted":false,"is_published":false,"raw_source_name":null,"raw_type":"article"},"sustainable_development_goals":[],"awards":[],"funders":[],"has_content":{"grobid_xml":false,"pdf":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Deep":[0],"ConvNets":[1],"suffer":[2],"from":[3,95],"gradient":[4,58,174],"signal":[5],"degradation":[6],"as":[7,170],"network":[8,41,125],"depth":[9],"increases,":[10],"limiting":[11],"effective":[12],"feature":[13],"learning":[14,183],"in":[15,184],"complex":[16],"architectures.":[17,187],"ResNet":[18],"addressed":[19],"this":[20],"through":[21,71],"residual":[22,54],"connections,":[23],"but":[24],"these":[25],"fixed":[26,53],"short-circuits":[27],"cannot":[28],"adapt":[29],"to":[30,100,116],"varying":[31],"input":[32],"complexity":[33],"or":[34],"selectively":[35],"emphasize":[36],"task":[37],"relevant":[38],"features":[39,64,78],"across":[40,91],"hierarchies.":[42],"This":[43],"study":[44],"introduces":[45],"GradAttn,":[46],"a":[47,177],"hybrid":[48],"CNN-transformer":[49],"framework":[50],"that":[51,105,131,145],"replaces":[52],"connections":[55],"with":[56,139,157],"attention-controlled":[57],"flow.":[59],"By":[60],"extracting":[61],"multi-scale":[62],"CNN":[63,158],"at":[65],"different":[66],"depths":[67],"and":[68,79],"regulating":[69],"them":[70],"self-attention,":[72],"GradAttn":[73,89,106],"dynamically":[74],"weights":[75],"shallow":[76],"texture":[77],"deep":[80,185],"semantic":[81],"representations.":[82],"For":[83],"representational":[84],"analysis,":[85],"we":[86],"evaluated":[87],"three":[88],"variants":[90],"eight":[92,112],"diverse":[93],"datasets,":[94,113],"natural":[96],"images,":[97],"medical":[98],"imaging,":[99],"fashion":[101],"recognition.":[102],"Results":[103],"demonstrate":[104],"outperforms":[107],"ResNet-18":[108],"on":[109,120],"five":[110],"of":[111,172],"achieving":[114],"up":[115],"+11.07%":[117],"accuracy":[118],"improvement":[119],"FashionMNIST":[121],"while":[122],"maintaining":[123],"comparable":[124],"size.":[126],"Gradient":[127],"flow":[128],"analysis":[129],"reveals":[130],"controlled":[132],"instabilities,":[133],"introduced":[134],"by":[135],"attention,":[136],"often":[137],"coincide":[138],"improved":[140],"generalization,":[141],"challenging":[142],"the":[143],"assumption":[144],"perfect":[146],"stability":[147],"is":[148],"optimal.":[149],"Furthermore,":[150],"positional":[151],"encoding":[152,161],"effectiveness":[153],"proves":[154],"dataset":[155],"dependent,":[156],"hierarchies":[159],"frequently":[160],"sufficient":[162],"spatial":[163],"structure.":[164],"These":[165],"findings":[166],"allow":[167],"attention":[168],"mechanisms":[169],"enablers":[171],"learnable":[173],"control,":[175],"offering":[176],"new":[178],"paradigm":[179],"for":[180],"adaptive":[181],"representation":[182],"neural":[186]},"counts_by_year":[],"updated_date":"2026-04-02T13:53:19.096889","created_date":"2026-04-02T00:00:00"}