{"id":"https://openalex.org/W7160412100","doi":"https://doi.org/10.48550/arxiv.2605.03969","title":"Feature-Augmented Transformers for Robust AI-Text Detection Across Domains and Generators","display_name":"Feature-Augmented Transformers for Robust AI-Text Detection Across Domains and Generators","publication_year":2026,"publication_date":"2026-05-05","ids":{"openalex":"https://openalex.org/W7160412100","doi":"https://doi.org/10.48550/arxiv.2605.03969"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2605.03969","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.03969","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.2605.03969","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5135514182","display_name":"Mohamed Mady","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Mady, Mohamed","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"middle","author":{"id":"https://openalex.org/A5135485321","display_name":"Johannes Reschke","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Reschke, Johannes","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5135486131","display_name":"Bjorn W Schuller","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Schuller, Bj\u00f6rn","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]}],"institutions":[],"countries_distinct_count":0,"institutions_distinct_count":0,"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":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":0.31450000405311584,"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":0.31450000405311584,"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/T11714","display_name":"Multimodal Machine Learning Applications","score":0.13860000669956207,"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":0.13600000739097595,"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/robustness","display_name":"Robustness (evolution)","score":0.7612000107765198},{"id":"https://openalex.org/keywords/detector","display_name":"Detector","score":0.5148000121116638},{"id":"https://openalex.org/keywords/transformer","display_name":"Transformer","score":0.4700999855995178},{"id":"https://openalex.org/keywords/vocabulary","display_name":"Vocabulary","score":0.42969998717308044},{"id":"https://openalex.org/keywords/readability","display_name":"Readability","score":0.40540000796318054},{"id":"https://openalex.org/keywords/feature","display_name":"Feature (linguistics)","score":0.34850001335144043},{"id":"https://openalex.org/keywords/boosting","display_name":"Boosting (machine learning)","score":0.33869999647140503}],"concepts":[{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.7612000107765198},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.6298999786376953},{"id":"https://openalex.org/C94915269","wikidata":"https://www.wikidata.org/wiki/Q1834857","display_name":"Detector","level":2,"score":0.5148000121116638},{"id":"https://openalex.org/C66322947","wikidata":"https://www.wikidata.org/wiki/Q11658","display_name":"Transformer","level":3,"score":0.4700999855995178},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.444599986076355},{"id":"https://openalex.org/C2777601683","wikidata":"https://www.wikidata.org/wiki/Q6499736","display_name":"Vocabulary","level":2,"score":0.42969998717308044},{"id":"https://openalex.org/C2778143727","wikidata":"https://www.wikidata.org/wiki/Q1820650","display_name":"Readability","level":2,"score":0.40540000796318054},{"id":"https://openalex.org/C2776401178","wikidata":"https://www.wikidata.org/wiki/Q12050496","display_name":"Feature (linguistics)","level":2,"score":0.34850001335144043},{"id":"https://openalex.org/C46686674","wikidata":"https://www.wikidata.org/wiki/Q466303","display_name":"Boosting (machine learning)","level":2,"score":0.33869999647140503},{"id":"https://openalex.org/C2779714256","wikidata":"https://www.wikidata.org/wiki/Q25305062","display_name":"Multiple Models","level":2,"score":0.3231000006198883},{"id":"https://openalex.org/C48372109","wikidata":"https://www.wikidata.org/wiki/Q3913","display_name":"Binary number","level":2,"score":0.31779998540878296},{"id":"https://openalex.org/C153180895","wikidata":"https://www.wikidata.org/wiki/Q7148389","display_name":"Pattern recognition (psychology)","level":2,"score":0.30869999527931213},{"id":"https://openalex.org/C148483581","wikidata":"https://www.wikidata.org/wiki/Q446488","display_name":"Feature selection","level":2,"score":0.29679998755455017},{"id":"https://openalex.org/C52622490","wikidata":"https://www.wikidata.org/wiki/Q1026626","display_name":"Feature extraction","level":2,"score":0.2913999855518341},{"id":"https://openalex.org/C11413529","wikidata":"https://www.wikidata.org/wiki/Q8366","display_name":"Algorithm","level":1,"score":0.29089999198913574},{"id":"https://openalex.org/C177774035","wikidata":"https://www.wikidata.org/wiki/Q1246948","display_name":"Granularity","level":2,"score":0.2791999876499176},{"id":"https://openalex.org/C149441793","wikidata":"https://www.wikidata.org/wiki/Q200726","display_name":"Probability distribution","level":2,"score":0.27090001106262207},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.2531000077724457}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2605.03969","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.03969","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.2605.03969","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.03969","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":{"AI-generated":[0],"text":[1],"is":[2,47,99],"nowadays":[3],"produced":[4],"at":[5],"scale":[6],"across":[7],"domains":[8],"and":[9,32,58,79,101,151,167,185],"heterogeneous":[10],"generation":[11],"pipelines,":[12],"making":[13],"robustness":[14,157],"to":[15,73,92,142,156],"distribution":[16],"shift":[17,98],"a":[18,34,168,182],"central":[19],"requirement":[20],"for":[21,51],"supervised":[22],"binary":[23],"detectors.":[24],"We":[25,64],"train":[26],"transformer-based":[27],"detectors":[28],"on":[29,42,67,80,122],"HC3":[30,68],"PLUS":[31],"calibrate":[33],"single":[35],"decision":[36],"threshold":[37,46],"by":[38,140],"maximising":[39],"balanced":[40,94,120],"accuracy":[41,121],"held-out":[43],"validation;":[44],"this":[45],"then":[48],"kept":[49],"fixed":[50],"all":[52],"downstream":[53],"test":[54],"distributions,":[55],"revealing":[56],"domain-":[57],"generator-dependent":[59],"error":[60],"asymmetries":[61],"under":[62,70,97,158],"shift.":[63,159],"evaluate":[65],"in-domain":[66,89],"PLUS,":[69],"cross-dataset":[71],"transfer":[72],"the":[74,81,130,178],"multi-domain,":[75],"multi-generator":[76],"M4":[77],"benchmark,":[78],"external":[82],"AI-Text-Detection-Pile.":[83],"Although":[84],"base":[85],"models":[86],"achieve":[87],"near-ceiling":[88],"performance":[90,96],"(up":[91],"99.5%":[93],"accuracy),":[95],"brittle":[100],"strongly":[102],"model-dependent.":[103],"Feature":[104],"augmentation":[105,166],"via":[106],"attention-based":[107],"linguistic":[108],"feature":[109,165],"fusion":[110],"improves":[111],"transfer,":[112],"with":[113],"our":[114,134],"best":[115],"model":[116,135],"(DeBERTa-v3-base+FeatAttn)":[117],"achieving":[118],"85.9%":[119],"M4.":[123],"Multi-seed":[124],"experiments":[125],"confirm":[126],"high":[127],"stability.":[128],"Under":[129],"same":[131],"fixed-threshold":[132,179],"protocol,":[133],"outperforms":[136],"strong":[137],"zero-shot":[138],"baselines":[139],"up":[141],"+7.22":[143],"points.":[144],"Category-level":[145],"ablations":[146],"further":[147],"show":[148],"that":[149,164],"readability":[150],"vocabulary":[152],"features":[153],"contribute":[154],"most":[155],"Overall,":[160],"these":[161],"results":[162],"demonstrate":[163],"modern":[169],"DeBERTa":[170],"backbone":[171],"significantly":[172],"outperform":[173],"earlier":[174],"BERT/RoBERTa":[175],"models,":[176],"while":[177],"protocol":[180],"provides":[181],"more":[183],"realistic":[184],"informative":[186],"assessment":[187],"of":[188],"practical":[189],"detector":[190],"robustness.":[191]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-05-07T00:00:00"}