DC FieldValueLanguage
dc.contributor.authorDjukic, Miodragen_US
dc.contributor.authorProkić, Ivanen_US
dc.contributor.authorPopovic, Miroslaven_US
dc.contributor.authorGhilezan, Silviaen_US
dc.contributor.authorPopovic, Markoen_US
dc.contributor.authorProkić, Simonaen_US
dc.date.accessioned2025-12-22T10:13:18Z-
dc.date.available2025-12-22T10:13:18Z-
dc.date.issued2025-
dc.identifier.issn1433-2779-
dc.identifier.urihttp://researchrepository.mi.sanu.ac.rs/handle/123456789/5636-
dc.description.abstractFederated learning (FL) is a machine learning setting where clients keep the training data decentralized and collaboratively train a model either under the coordination of a central server (centralized FL) or in a peer-to-peer network (decentralized FL). Correct orchestration is one of the main challenges. In this paper, we formally verify the correctness of two generic FL algorithms, a centralized and a decentralized one, using the Communicating Sequential Processes (CSP) calculus and the Process Analysis Toolkit (PAT) model checker. The CSP models consist of CSP processes corresponding to generic FL algorithm instances. PAT automatically proves the correctness of the two generic FL algorithms by proving their deadlock freedom (safety property) and successful termination (reachability and liveness property). The CSP models are constructed as a faithful representation of the real Python code and are expressed directly in CSP# language that PAT uses. Then they are automatically checked top-down by PAT. The Python code follows a restricted actor-based programming model, and the construction of CSP# code from such Python code is performed systematically. The process is described in detail, ensuring that the models correspond to the actual code. It represents a basis for developing tools for automatic translation of certain classes of Python code to CSP models, expressed in CSP#.en_US
dc.publisherSpringer Linken_US
dc.relationThis work was supported by Horizon EU project 101093006 TaRDIS, MSTDI (Contract No. 451-03-137/2025-03/200156) and FTS-UNS through the project Scientific and Artistic Research Work of Researchers in Teaching and Associate Positions at FTS-UNS 2025 (No. 01-50/295), COST CA20111 EuroProofNet, and EPSRC grants EP/T006544/2, EP/N027833/2,EP/T014709/2, EP/V000462/1, EP/X015955/1, EP/Y005244/1 and ARIA.en_US
dc.relation.ispartofInternational Journal on Software Tools for Technology Transferen_US
dc.subjectActor model | CSP process calculus | Decentralized intelligence | Federated learning | Formal verification | PAT model checker | Pythonen_US
dc.titleCorrect orchestration of federated learning generic algorithms: Python translation to CSP and verification by PATen_US
dc.typeArticleen_US
dc.identifier.doi10.1007/s10009-025-00795-0-
dc.identifier.scopus2-s2.0-105003740675-
dc.contributor.affiliationMathematicsen_US
dc.contributor.affiliationMathematical Institute of the Serbian Academy of Sciences and Artsen_US
dc.relation.firstpage21-
dc.relation.lastpage34-
dc.relation.volume27-
dc.description.rank~M22-
item.fulltextNo Fulltext-
item.cerifentitytypePublications-
item.grantfulltextnone-
item.openairetypeArticle-
item.openairecristypehttp://purl.org/coar/resource_type/c_18cf-
crisitem.author.orcid0000-0003-2253-8285-
Show simple item record

SCOPUSTM   
Citations

3
checked on Feb 11, 2026

Page view(s)

48
checked on Feb 13, 2026

Google ScholarTM

Check

Altmetric

Altmetric


Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.