Briefing
The core research problem is the secure, private, and publicly verifiable aggregation of data from numerous clients by an untrusted server in a non-interactive setting. The paper introduces the Non-Interactive Verifiable Aggregation (NIVA) primitive, realized by the PEAR protocol, which achieves this by seamlessly integrating Inner-Product Functional Encryption for client privacy with fully-linear Probabilistically-Checkable Proofs for public verifiability. This mechanism allows a “weak” analyst to obtain an aggregate result without seeing individual client inputs, while simultaneously ensuring that malicious clients cannot skew the final statistic. This breakthrough establishes a foundational building block for future decentralized applications that require massive-scale, trust-minimized, and privacy-preserving data collection, such as secure federated learning and anonymous on-chain voting.
Context
Prior to NIVA, verifiable secure aggregation protocols faced a fundamental trade-off between non-interactivity and robustness. Existing solutions either required clients to remain online for multiple rounds (interactive), relied on multiple non-colluding servers (decentralized server model), or lacked a mechanism for a public third party to verify the final aggregate result. The prevailing theoretical limitation was the difficulty of cryptographically guaranteeing both client-side input validity (robustness against malicious clients) and data privacy against a single, powerful, but untrusted aggregation server.
Analysis
The core mechanism is the PEAR protocol, which instantiates the NIVA primitive by leveraging two advanced cryptographic tools in a black-box manner. First, clients use Inner-Product Functional Encryption (IPFE) to encrypt their secret input vector, ensuring the server can only compute the inner product (the aggregate sum) and nothing else. Second, the client attaches a non-interactive proof, built from fully-linear Probabilistically-Checkable Proofs (PCPs), that mathematically certifies their encrypted input is valid according to a pre-defined NP validity rule.
The server performs the aggregation on the ciphertexts, and the analyst verifies the aggregate result using the proofs. This structure fundamentally differs from previous work by making the verification public and non-interactive, eliminating the need for client presence after the initial submission and replacing trust in the server with mathematical proof.
Parameters
- Non-Interactive, Public-Key, Single-Server → The NIVA primitive is the first to simultaneously satisfy all three properties for aggregation with input validation.
- Core Cryptographic Tool → Inner-Product Functional Encryption → Used to enable the server to compute the aggregate sum on encrypted data without learning individual inputs.
- Proof System Basis → Fully-Linear Probabilistically-Checkable Proofs → Provides the cryptographic foundation for the non-interactive, publicly verifiable input validity check.
Outlook
This research opens new avenues for decentralized systems by decoupling the privacy-preserving computation layer from the trust model. Future work will focus on improving the efficiency of the underlying unrestricted IPFE schemes, which is the current bottleneck, to unlock greater performance gains. The NIVA primitive is a crucial step toward enabling practical, large-scale, privacy-preserving applications such as auditable decentralized governance systems, secure medical data pooling, and robust, non-Byzantine federated machine learning in Web3 environments within the next five years.
Verdict
The Non-Interactive Verifiable Aggregation primitive is a fundamental cryptographic advance that resolves the core trade-off between privacy, verifiability, and non-interactivity in secure data aggregation protocols.
