✅ Funded — ANR (JCJC, 2026 - 2030)

ACAMAS

Advanced Confidentiality and Authenticity for Multi-Agent Systems

FundingAgence Nationale de la Recherche (ANR), AAPG 2026 (Jeune Chercheur-Jeune Chercheuse (JCJC) scheme for young researchers)
Duration48 months (October 2026 - September 2030)
Host institutionSorbonne Université, team ALMASTY, LIP6 (UMR 7606)
Principal InvestigatorKy Nguyen (Associate Professor, Sorbonne Université)
Research CollaboratorsMichael Reichle (Tenure-track Faculty Member, INSAIT) ; Erkan Tairi (Postdoctoral Researcher, UC Berkeley)

Overview

Modern communication environments involve complex interactions between entities with divergent interests, heterogeneous capabilities, and individual data, whether in web content distribution systems, federated learning, or authentication for electronic payments. Despite the strict constraints of the GDPR on personal data processing, emerging threats in decentralized computing demand advanced cryptographic solutions that ensure both privacy protection and fine-grained data analysis. ACAMAS aims to develop advanced cryptographic solutions for multi-agent systems, unifying two complementary aspects: data confidentiality, e.g., through multi-client functional encryption, and privacy-preserving authentication, e.g., through blind and threshold signatures. The approach relies on decentralization and the removal of single points of failure in complex environments where multiple agents (users, receivers, authorities) may be corrupted.

Main Challenges

Objectives

Multi-Agent Functional Encryption: Dynamic, Decentralized, Post-Quantum

Strengthen security models for decentralized and dynamic multi-client functional encryption (DMCFE/DDFE) with fine-grained access control, then push these constructions toward post-quantum security — first via lattice-based techniques, and exploring code-based and isogeny-based alternatives where no analogue currently exists.

Active Security and Privacy-Preserving Authentication

Extend chosen-ciphertext (CCA) security to the multi-agent functional encryption setting and investigate threshold decryption, while advancing threshold and blind signatures, all with an eye toward concrete contributions to NIST's multi-party threshold standardization process.

Related Publications Talks & Dissemination
Back to Homepage