{"schema_version":"0.1","map_id":"paper-21-map","publication_id":21,"publication_anchor":"paper-21","slug":"paper-21","canonical_path":"/knowledge/papers/paper-21/","machine_path":"/knowledge/papers/paper-21.json","root_node_id":"paper-21","stage":"mapped_draft","contribution_type_vocabulary_version":"0.1","contribution_types":["protocol","primitive"],"title":"5PM: Secure Pattern Matching","short_title":"5PM","year":2013,"status":"Published · journal article","venue":"Journal of Computer Security, Volume 21, Number 5","topic":"secure-encrypted-computation","labels":["Theory"],"authors":["Joshua Baron","Karim Eldefrawy","Kirill Minkovich","Rafail Ostrovsky","Eric Tressler"],"keywords":["secure pattern matching","privacy-preserving search"],"research_question":"Can two parties securely evaluate expressive pattern matching over a client's private pattern and a server's private text, including nonbinary alphabets, wildcards, substring/Hamming-distance matching, and optional pattern-length hiding, with malicious security and communication sublinear in the corresponding circuit size?","central_answer":"5PM converts pattern matching into linear operations over text and character-delay matrices, evaluates those operations with additively homomorphic encryption, and adds threshold encryption plus zero-knowledge consistency arguments for static malicious security; the paper proves two-round honest-but-curious and eight-round malicious variants and reports an implementation of the honest-but-curious variant.","curation":{"drafted_at":"2026-07-11","drafted_by":[{"actor_type":"ai","name":"OpenAI Codex","role":"full-text extraction, claim decomposition, evidence linking, and initial assessment"}],"method":"Source-grounded review of the complete 51-page UCLA author-hosted version, including protocol definitions, security proofs, and implementation results; PDF pages 1 and 12 were also rendered and visually inspected. The final journal pagination and text were not compared line by line with this author version.","source_scope":"full_source_audit","approval":{"status":"pending","note":"AI-authored source-linked map awaiting author verification. Formal claims were transcribed conservatively and have not been independently reproved or reproduced."}},"sources":[{"id":"source-paper-21-fulltext","type":"scholarly_article","title":"5PM: Secure Pattern Matching (author-hosted full version)","url":"/pubs/2013/5pm-secure-pattern-matching.pdf","media_type":"application/pdf","sha256":"e7628754d71352f7bae8ae8097200db10912c86e7c15d85c9627e63cc53df7e6","page_count":51,"provenance_category":"author","version_note":"Mirrors the public UCLA author copy; correspondence with the final journal typesetting was not checked line by line."},{"id":"source-paper-21-author-origin","type":"author_copy","title":"UCLA author copy","url":"https://web.cs.ucla.edu/~rafail/PUBLIC/129.pdf"},{"id":"source-paper-21-archive","type":"manuscript_archive","title":"IACR ePrint 2012/698","url":"https://eprint.iacr.org/2012/698"},{"id":"source-paper-21-official","type":"publication_record","title":"Journal of Computer Security publication record","url":"https://doi.org/10.3233/JCS-130481"},{"id":"source-paper-21-citations","type":"scholarly_index","title":"OpenAlex work record for paper #21","url":"https://openalex.org/W1806693672","accessed_at":"2026-07-11"}],"source_anchors":[{"id":"anchor-paper-21-problem","source_id":"source-paper-21-fulltext","label":"Problem, functionality, complexity, and round claims","locator":"Abstract and Section 1.1, PDF pages 1-3","url":"/pubs/2013/5pm-secure-pattern-matching.pdf#page=1"},{"id":"anchor-paper-21-input-output","source_id":"source-paper-21-fulltext","label":"Pattern-matching inputs, outputs, wildcards, and substring semantics","locator":"Sections 1 and 2.1, PDF pages 1-7","url":"/pubs/2013/5pm-secure-pattern-matching.pdf#page=1"},{"id":"anchor-paper-21-tools","source_id":"source-paper-21-fulltext","label":"Cryptographic tools and linear-algebra operators","locator":"Sections 2.2-2.3, PDF pages 7-10","url":"/pubs/2013/5pm-secure-pattern-matching.pdf#page=7"},{"id":"anchor-paper-21-hbc","source_id":"source-paper-21-fulltext","label":"Honest-but-curious protocol and Theorem 1","locator":"Sections 3.1-3.2, PDF pages 10-12","url":"/pubs/2013/5pm-secure-pattern-matching.pdf#page=10"},{"id":"anchor-paper-21-malicious","source_id":"source-paper-21-fulltext","label":"Eight-round malicious protocol and consistency subprotocols","locator":"Section 3.3, PDF pages 12-17; exact interleaving in Section 6","url":"/pubs/2013/5pm-secure-pattern-matching.pdf#page=12"},{"id":"anchor-paper-21-security-model","source_id":"source-paper-21-fulltext","label":"Ideal functionality, static corruption model, and simulators","locator":"Section 7, PDF pages 29-44","url":"/pubs/2013/5pm-secure-pattern-matching.pdf#page=29"},{"id":"anchor-paper-21-security-theorems","source_id":"source-paper-21-fulltext","label":"HBC and malicious realization theorems","locator":"Theorems 4 and 5, PDF pages 31 and 37","url":"/pubs/2013/5pm-secure-pattern-matching.pdf#page=31"},{"id":"anchor-paper-21-performance","source_id":"source-paper-21-fulltext","label":"Implementation platform and timing table","locator":"Section 8 and Table 13, PDF pages 48-49","url":"/pubs/2013/5pm-secure-pattern-matching.pdf#page=48"},{"id":"anchor-paper-21-publication","source_id":"source-paper-21-official","label":"Official journal record","locator":"Journal of Computer Security, volume 21, number 5","url":"https://doi.org/10.3233/JCS-130481"},{"id":"anchor-paper-21-citations","source_id":"source-paper-21-citations","label":"Dated OpenAlex citation snapshot","locator":"cited_by_count = 17, accessed 2026-07-11","url":"https://openalex.org/W1806693672"}],"nodes":[{"id":"paper-21","kind":"paper","parent_id":null,"order":1,"epistemic_status":"published","title":"5PM: Secure Pattern Matching","summary":"A two-party secure pattern-matching protocol family for exact matching, single-character wildcards, nonbinary Hamming distance, and substring matching, with explicit honest-but-curious and static-malicious constructions.","source_anchor_ids":["anchor-paper-21-problem"]},{"id":"paper-21-question","kind":"question","parent_id":"paper-21","order":1,"epistemic_status":"research_question","title":"Research question","summary":"Can expressive private pattern matching obtain malicious security, bounded rounds, O(m+n) ciphertext-scale communication, and optional hiding of the pattern length?","source_anchor_ids":["anchor-paper-21-problem"]},{"id":"paper-21-answer","kind":"contribution","parent_id":"paper-21","order":2,"epistemic_status":"source_asserted","title":"Central answer","summary":"Encode the server text and client pattern as matrices, reduce matching to linear operators, evaluate one encrypted operand homomorphically, and enforce well-formedness and consistent computation with threshold decryption and zero-knowledge arguments in the malicious variant.","source_anchor_ids":["anchor-paper-21-tools","anchor-paper-21-malicious"]},{"id":"paper-21-scope","kind":"scope","parent_id":"paper-21","order":3,"epistemic_status":"defined","title":"Functionality, parties, and adversary","summary":"Server holds text T of length n over alphabet Sigma; Client holds pattern p of length m, optionally containing wildcards, and learns either matching locations or a decision while Server learns no output.","source_anchor_ids":["anchor-paper-21-input-output","anchor-paper-21-security-model"]},{"id":"paper-21-scope-functionality","kind":"definition","parent_id":"paper-21-scope","order":1,"epistemic_status":"defined","title":"Matching semantics","summary":"The same activation-vector construction supports exact matches, single-character wildcards, nonbinary Hamming distance, and thresholded substring matching; multiple distance thresholds can reuse one protocol execution.","source_anchor_ids":["anchor-paper-21-input-output","anchor-paper-21-hbc"]},{"id":"paper-21-scope-adversary","kind":"threat_model","parent_id":"paper-21-scope","order":2,"epistemic_status":"defined","title":"Static two-party corruption","summary":"The formal analyses treat one statically corrupted party in either the honest-but-curious or malicious stand-alone setting; the malicious proof is simulation-based for the stated ideal functionality.","source_anchor_ids":["anchor-paper-21-security-model","anchor-paper-21-security-theorems"]},{"id":"paper-21-scope-assumptions","kind":"assumption","parent_id":"paper-21-scope","order":3,"epistemic_status":"assumed","title":"Cryptographic assumptions","summary":"The HBC theorem assumes semantically secure additive homomorphic encryption over a prime-order cyclic group; the concrete malicious construction uses threshold ElGamal and assumes DDH hardness plus the specified commitment and zero-knowledge machinery.","source_anchor_ids":["anchor-paper-21-tools","anchor-paper-21-security-theorems"]},{"id":"paper-21-method","kind":"method","parent_id":"paper-21","order":4,"epistemic_status":"specified","title":"Construction path","summary":"5PM transforms the insecure character-delay-vector algorithm into matrix multiplication, Stretch, Cut, and column-sum operations that remain computable when one operand is encrypted.","source_anchor_ids":["anchor-paper-21-input-output","anchor-paper-21-tools"]},{"id":"paper-21-method-hbc","kind":"component","parent_id":"paper-21-method","order":1,"epistemic_status":"specified_and_proved","title":"Two-round HBC protocol","summary":"Client encrypts its character-delay matrix and threshold; Server forms and blinds an encrypted activation vector; Client decrypts zero entries to identify matches.","source_anchor_ids":["anchor-paper-21-hbc"]},{"id":"paper-21-method-malicious","kind":"component","parent_id":"paper-21-method","order":2,"epistemic_status":"specified_and_proved","title":"Eight-round malicious protocol","summary":"Both parties independently derive encrypted activation vectors, compare affine hashes, and prove matrix formation, partial decryption, randomization, and final-decryption consistency through interleaved zero-knowledge arguments.","source_anchor_ids":["anchor-paper-21-malicious"]},{"id":"paper-21-claims","kind":"claim_group","parent_id":"paper-21","order":5,"epistemic_status":"mixed","title":"Principal claims","summary":"The paper makes functionality, round/complexity, simulation-security, and implementation-performance claims.","source_anchor_ids":["anchor-paper-21-problem","anchor-paper-21-security-theorems","anchor-paper-21-performance"]},{"id":"paper-21-claim-complexity","kind":"claim","parent_id":"paper-21-claims","order":1,"epistemic_status":"analytically_supported","title":"Rounds and asymptotic cost","summary":"The paper reports two one-way rounds for HBC and eight for malicious security; the malicious construction uses O((m+n)k^2) bandwidth and O(m+n) encryptions, while pattern-length hiding adds no asymptotic computation or bandwidth.","source_anchor_ids":["anchor-paper-21-problem","anchor-paper-21-malicious"]},{"id":"paper-21-claim-security","kind":"claim","parent_id":"paper-21-claims","order":2,"epistemic_status":"proved_conditional","title":"Simulation security","summary":"Theorems 4 and 5 state realization of the specified pattern-matching functionality for static HBC corruption under semantic security and static malicious corruption under DDH, respectively.","source_anchor_ids":["anchor-paper-21-security-theorems"]},{"id":"paper-21-claim-implementation","kind":"claim","parent_id":"paper-21-claims","order":3,"epistemic_status":"experimentally_supported","title":"HBC implementation","summary":"A Paillier-based HBC prototype is timed on one dual-quad-core 2.93 GHz Ubuntu 10.10 machine for DNA and alphanumeric alphabets, text lengths through 100,000, and 1024/2048-bit keys.","source_anchor_ids":["anchor-paper-21-performance"]},{"id":"paper-21-evidence","kind":"evidence_group","parent_id":"paper-21","order":6,"epistemic_status":"documented","title":"Evidence chain","summary":"Support consists of explicit protocols, ideal-world simulations, asymptotic accounting, and a timing table for an HBC prototype.","source_anchor_ids":["anchor-paper-21-security-model","anchor-paper-21-performance"]},{"id":"paper-21-evidence-formal","kind":"evidence","parent_id":"paper-21-evidence","order":1,"epistemic_status":"paper_proof_not_machine_checked","title":"Formal proof path","summary":"The full version specifies ideal functionalities and simulators for corrupted Client and Server; this map records the theorem boundaries but has not mechanically checked every hybrid and extractor argument.","source_anchor_ids":["anchor-paper-21-security-model","anchor-paper-21-security-theorems"]},{"id":"paper-21-evidence-empirical","kind":"evidence","parent_id":"paper-21-evidence","order":2,"epistemic_status":"reported_not_reproduced","title":"Benchmark evidence","summary":"Table 13 decomposes search, blinding, and decryption time across parameter choices, but supplies neither source code nor repeated-run uncertainty in the audited full version.","source_anchor_ids":["anchor-paper-21-performance"]},{"id":"paper-21-boundaries","kind":"limitation_group","parent_id":"paper-21","order":7,"epistemic_status":"material","title":"Boundaries and limitations","summary":"The strongest statements are conditional on a static two-party model and cryptographic assumptions; empirical evidence covers only the HBC prototype.","source_anchor_ids":["anchor-paper-21-security-model","anchor-paper-21-performance"]},{"id":"paper-21-boundary-model","kind":"limitation","parent_id":"paper-21-boundaries","order":1,"epistemic_status":"out_of_scope","title":"Unmodeled deployment risks","summary":"The proofs do not cover adaptive corruption, multi-session/composable deployment, leakage through repeated queries, implementation side channels, key compromise, or denial of service.","source_anchor_ids":["anchor-paper-21-security-model"]},{"id":"paper-21-boundary-empirical","kind":"limitation","parent_id":"paper-21-boundaries","order":2,"epistemic_status":"explicitly_limited","title":"No malicious-protocol benchmark","summary":"The performance section evaluates the honest-but-curious implementation; it does not report an implementation or benchmark of the eight-round malicious protocol.","source_anchor_ids":["anchor-paper-21-performance"]},{"id":"paper-21-artifacts","kind":"artifact_group","parent_id":"paper-21","order":8,"epistemic_status":"full_text_available","title":"Artifacts and resources","summary":"A fixed local mirror of the 51-page author version, its public UCLA origin, an IACR ePrint record, and the journal DOI are available; no code repository was identified in the audited paper.","source_anchor_ids":["anchor-paper-21-problem","anchor-paper-21-publication"]},{"id":"paper-21-scrutiny","kind":"scrutiny","parent_id":"paper-21","order":9,"epistemic_status":"publication_recorded","title":"External scrutiny","summary":"The work has a Journal of Computer Security publication record; review reports, correction history, independent reproduction, and later cryptanalysis were not audited.","source_anchor_ids":["anchor-paper-21-publication"]},{"id":"paper-21-lineage","kind":"lineage","parent_id":"paper-21","order":10,"epistemic_status":"source_asserted","title":"Research lineage","summary":"5PM extends secure exact matching toward nonbinary wildcards and substring/Hamming-distance functionality by exploiting a linear-algebra formulation rather than generic circuit evaluation.","source_anchor_ids":["anchor-paper-21-problem","anchor-paper-21-tools"]}],"relations":[{"id":"relation-paper-21-answer-addresses-question","type":"addresses","from_id":"paper-21-answer","to_id":"paper-21-question"},{"id":"relation-paper-21-method-realizes-answer","type":"realizes","from_id":"paper-21-method","to_id":"paper-21-answer"},{"id":"relation-paper-21-formal-supports-security","type":"supports","from_id":"paper-21-evidence-formal","to_id":"paper-21-claim-security"},{"id":"relation-paper-21-empirical-supports-implementation","type":"supports","from_id":"paper-21-evidence-empirical","to_id":"paper-21-claim-implementation"},{"id":"relation-paper-21-assumptions-qualify-security","type":"qualifies","from_id":"paper-21-scope-assumptions","to_id":"paper-21-claim-security"},{"id":"relation-paper-21-model-qualifies-security","type":"qualifies","from_id":"paper-21-boundary-model","to_id":"paper-21-claim-security"},{"id":"relation-paper-21-empirical-boundary-qualifies-implementation","type":"qualifies","from_id":"paper-21-boundary-empirical","to_id":"paper-21-claim-implementation"},{"id":"relation-paper-21-artifacts-enable-audit","type":"enables_audit_of","from_id":"paper-21-artifacts","to_id":"paper-21-evidence"}],"assessment":{"id":"paper-21-assessment-2026-07-11","rubric_version":"0.2","assessed_at":"2026-07-11","status":"ai_draft_author_review_pending","note":"These dimensions describe documented support and process, not truth, correctness, or a universal ranking. No composite score is calculated.","axes":[{"id":"epistemic_evidence","level":"medium","rationale":"The source gives complete protocol specifications, simulation arguments, asymptotic analysis, and HBC timing measurements. The proofs were not mechanically checked and the implementation evidence was not reproduced; the malicious construction was not benchmarked.","basis_source_anchor_ids":["anchor-paper-21-security-theorems","anchor-paper-21-performance"]},{"id":"auditability","level":"high","rationale":"A complete author-hosted full version is mirrored locally with page count and SHA-256, so assumptions, protocols, proofs, and reported measurements are directly inspectable. Journal-version identity and empirical reproduction remain unchecked.","basis_source_anchor_ids":["anchor-paper-21-problem","anchor-paper-21-publication"]},{"id":"production_provenance","level":"medium","rationale":"Named authorship, funding acknowledgements, an author-hosted version, archive record, and journal record establish a publication trail. Contributor roles, revision history, code lineage, and production effort are not documented in this map.","basis_source_anchor_ids":["anchor-paper-21-problem","anchor-paper-21-publication"]},{"id":"external_scrutiny","level":"medium","rationale":"The paper has a journal publication record, but reviews, independent replications, later attacks, and correction history were not audited.","basis_source_anchor_ids":["anchor-paper-21-publication"]},{"id":"reception","level":"high","rationale":"OpenAlex reports 17 located citations as of 2026-07-11, meeting the site's 11+ high threshold. The count is index-specific and does not itself establish correctness or adoption.","basis_source_anchor_ids":["anchor-paper-21-citations"]},{"id":"contribution_significance","level":"medium","rationale":"The paper presents a specific protocol family with expressive functionality and explicit malicious security/round claims; priority and field-level impact were not independently evaluated.","basis_source_anchor_ids":["anchor-paper-21-problem"]}]},"reception_snapshot":{"as_of":"2026-07-11","method":"OpenAlex DOI work record cited_by_count","source_url":"https://openalex.org/W1806693672","citation_count":17,"signals":[],"limitation":"OpenAlex coverage, deduplication, and version merging are imperfect; this is a dated located-citation snapshot."}}