{"id":"https://openalex.org/W7160996006","doi":"https://doi.org/10.48550/arxiv.2605.10977","title":"PASA: A Principled Embedding-Space Watermarking Approach for LLM-Generated Text under Semantic-Invariant Attacks","display_name":"PASA: A Principled Embedding-Space Watermarking Approach for LLM-Generated Text under Semantic-Invariant Attacks","publication_year":2026,"publication_date":"2026-05-09","ids":{"openalex":"https://openalex.org/W7160996006","doi":"https://doi.org/10.48550/arxiv.2605.10977"},"language":null,"primary_location":{"id":"doi:10.48550/arxiv.2605.10977","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.10977","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.2605.10977","any_repository_has_fulltext":true},"authorships":[{"author_position":"first","author":{"id":"https://openalex.org/A5116834253","display_name":"Zhenxin Ai","orcid":"https://orcid.org/0009-0007-4720-2442"},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"Ai, Zhenxin","raw_affiliation_strings":[],"raw_orcid":null,"affiliations":[]},{"author_position":"last","author":{"id":"https://openalex.org/A5136067417","display_name":"Haiyun He","orcid":null},"institutions":[],"countries":[],"is_corresponding":false,"raw_author_name":"He, Haiyun","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/T10388","display_name":"Advanced Steganography and Watermarking Techniques","score":0.3797999918460846,"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/T10388","display_name":"Advanced Steganography and Watermarking Techniques","score":0.3797999918460846,"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/T11689","display_name":"Adversarial Robustness in Machine Learning","score":0.33000001311302185,"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/T11241","display_name":"Advanced Malware Detection Techniques","score":0.03099999949336052,"subfield":{"id":"https://openalex.org/subfields/1711","display_name":"Signal Processing"},"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/digital-watermarking","display_name":"Digital watermarking","score":0.8310999870300293},{"id":"https://openalex.org/keywords/key","display_name":"Key (lock)","score":0.6378999948501587},{"id":"https://openalex.org/keywords/embedding","display_name":"Embedding","score":0.5633000135421753},{"id":"https://openalex.org/keywords/randomness","display_name":"Randomness","score":0.5464000105857849},{"id":"https://openalex.org/keywords/watermark","display_name":"Watermark","score":0.4941999912261963},{"id":"https://openalex.org/keywords/semantics","display_name":"Semantics (computer science)","score":0.4896000027656555},{"id":"https://openalex.org/keywords/dependency","display_name":"Dependency (UML)","score":0.40220001339912415},{"id":"https://openalex.org/keywords/security-token","display_name":"Security token","score":0.3995000123977661}],"concepts":[{"id":"https://openalex.org/C150817343","wikidata":"https://www.wikidata.org/wiki/Q875932","display_name":"Digital watermarking","level":3,"score":0.8310999870300293},{"id":"https://openalex.org/C41008148","wikidata":"https://www.wikidata.org/wiki/Q21198","display_name":"Computer science","level":0,"score":0.7376999855041504},{"id":"https://openalex.org/C26517878","wikidata":"https://www.wikidata.org/wiki/Q228039","display_name":"Key (lock)","level":2,"score":0.6378999948501587},{"id":"https://openalex.org/C41608201","wikidata":"https://www.wikidata.org/wiki/Q980509","display_name":"Embedding","level":2,"score":0.5633000135421753},{"id":"https://openalex.org/C125112378","wikidata":"https://www.wikidata.org/wiki/Q176640","display_name":"Randomness","level":2,"score":0.5464000105857849},{"id":"https://openalex.org/C164112704","wikidata":"https://www.wikidata.org/wiki/Q7974348","display_name":"Watermark","level":3,"score":0.4941999912261963},{"id":"https://openalex.org/C184337299","wikidata":"https://www.wikidata.org/wiki/Q1437428","display_name":"Semantics (computer science)","level":2,"score":0.4896000027656555},{"id":"https://openalex.org/C154945302","wikidata":"https://www.wikidata.org/wiki/Q11660","display_name":"Artificial intelligence","level":1,"score":0.4691999852657318},{"id":"https://openalex.org/C19768560","wikidata":"https://www.wikidata.org/wiki/Q320727","display_name":"Dependency (UML)","level":2,"score":0.40220001339912415},{"id":"https://openalex.org/C48145219","wikidata":"https://www.wikidata.org/wiki/Q1335365","display_name":"Security token","level":2,"score":0.3995000123977661},{"id":"https://openalex.org/C148417208","wikidata":"https://www.wikidata.org/wiki/Q4825882","display_name":"Authentication (law)","level":2,"score":0.3991999924182892},{"id":"https://openalex.org/C77618280","wikidata":"https://www.wikidata.org/wiki/Q1155772","display_name":"Scheme (mathematics)","level":2,"score":0.3806000053882599},{"id":"https://openalex.org/C80444323","wikidata":"https://www.wikidata.org/wiki/Q2878974","display_name":"Theoretical computer science","level":1,"score":0.36390000581741333},{"id":"https://openalex.org/C3073032","wikidata":"https://www.wikidata.org/wiki/Q15912075","display_name":"Information hiding","level":3,"score":0.36230000853538513},{"id":"https://openalex.org/C204321447","wikidata":"https://www.wikidata.org/wiki/Q30642","display_name":"Natural language processing","level":1,"score":0.3452000021934509},{"id":"https://openalex.org/C25810664","wikidata":"https://www.wikidata.org/wiki/Q44325","display_name":"Ontology","level":2,"score":0.34389999508857727},{"id":"https://openalex.org/C99138194","wikidata":"https://www.wikidata.org/wiki/Q183427","display_name":"Hash function","level":2,"score":0.33149999380111694},{"id":"https://openalex.org/C55439883","wikidata":"https://www.wikidata.org/wiki/Q360812","display_name":"Correctness","level":2,"score":0.32989999651908875},{"id":"https://openalex.org/C115961682","wikidata":"https://www.wikidata.org/wiki/Q860623","display_name":"Image (mathematics)","level":2,"score":0.32679998874664307},{"id":"https://openalex.org/C2778572836","wikidata":"https://www.wikidata.org/wiki/Q380933","display_name":"Space (punctuation)","level":2,"score":0.2985000014305115},{"id":"https://openalex.org/C124101348","wikidata":"https://www.wikidata.org/wiki/Q172491","display_name":"Data mining","level":1,"score":0.29840001463890076},{"id":"https://openalex.org/C178489894","wikidata":"https://www.wikidata.org/wiki/Q8789","display_name":"Cryptography","level":2,"score":0.2865999937057495},{"id":"https://openalex.org/C23123220","wikidata":"https://www.wikidata.org/wiki/Q816826","display_name":"Information retrieval","level":1,"score":0.2721000015735626},{"id":"https://openalex.org/C63479239","wikidata":"https://www.wikidata.org/wiki/Q7353546","display_name":"Robustness (evolution)","level":3,"score":0.271699994802475},{"id":"https://openalex.org/C119857082","wikidata":"https://www.wikidata.org/wiki/Q2539","display_name":"Machine learning","level":1,"score":0.2678000032901764},{"id":"https://openalex.org/C203062551","wikidata":"https://www.wikidata.org/wiki/Q201339","display_name":"Public-key cryptography","level":3,"score":0.25040000677108765}],"mesh":[],"locations_count":1,"locations":[{"id":"doi:10.48550/arxiv.2605.10977","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.10977","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.2605.10977","is_oa":true,"landing_page_url":"https://doi.org/10.48550/arxiv.2605.10977","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":[{"score":0.6660643815994263,"display_name":"Quality Education","id":"https://metadata.un.org/sdg/4"}],"awards":[],"funders":[],"has_content":{"pdf":false,"grobid_xml":false},"content_urls":null,"referenced_works_count":0,"referenced_works":[],"related_works":[],"abstract_inverted_index":{"Watermarking":[0],"for":[1,10],"large":[2],"language":[3],"models":[4],"(LLMs)":[5],"is":[6,84],"a":[7,34,45,57,63,76,92],"promising":[8],"approach":[9],"detecting":[11],"LLM-generated":[12],"text":[13,127],"and":[14,37,43,61,68,79,105,111],"enabling":[15],"responsible":[16],"deployment.":[17],"However,":[18],"existing":[19],"watermarking":[20,39],"methods":[21],"are":[22],"often":[23],"vulnerable":[24],"to":[25],"semantic-invariant":[26,112],"attacks,":[27],"such":[28],"as":[29],"paraphrasing.":[30],"We":[31],"propose":[32],"PASA,":[33],"principled,":[35],"robust,":[36],"distortion-free":[38],"algorithm":[40],"that":[41,90,115],"embeds":[42],"detects":[44],"watermark":[46],"at":[47],"the":[48,98,137],"semantic":[49,54,80],"level.":[50],"PASA":[51,116],"operates":[52],"on":[53],"clusters":[55],"in":[56,86],"latent":[58],"embedding":[59],"space":[60],"constructs":[62],"distributional":[64],"dependency":[65],"between":[66],"token":[67],"auxiliary":[69],"sequences":[70],"via":[71],"shared":[72],"randomness":[73],"synchronized":[74],"by":[75],"secret":[77],"key":[78],"history.":[81],"This":[82],"design":[83],"grounded":[85],"our":[87,140],"theoretical":[88],"framework":[89],"characterizes":[91],"jointly":[93],"optimal":[94],"embedding-detection":[95],"pair,":[96],"achieving":[97],"fundamental":[99],"trade-offs":[100],"among":[101],"detection":[102],"accuracy,":[103],"robustness,":[104],"distortion.":[106],"Evaluations":[107],"across":[108],"multiple":[109],"LLMs":[110],"attacks":[113,123],"demonstrate":[114],"remains":[117],"robust":[118],"even":[119],"under":[120],"strong":[121],"paraphrasing":[122],"while":[124],"preserving":[125],"high":[126],"quality,":[128],"outperforming":[129],"standard":[130],"vocabulary-space":[131],"baselines.":[132],"Ablation":[133],"studies":[134],"further":[135],"validate":[136],"effectiveness":[138],"of":[139],"hyperparameter":[141],"choices.":[142],"Webpage:":[143],"https://ai-kunkun.github.io/PASA_page/.":[144]},"counts_by_year":[],"updated_date":"2026-07-01T06:00:48.157686","created_date":"2026-05-14T00:00:00"}