Program Summary
A zero-knowledge (ZK) proof is an interactive protocol between a prover and a verifier. The prover creates a statement that they want the verifier to accept, using knowledge that will remain hidden from the verifier. Recent research has substantially increased the efficiency of ZK proofs, enabling real-world use, primarily by cryptocurrencies. While useful for cryptocurrencies, the ZK proofs created are specialized for this task and do not necessarily scale for transactions that are more complex. For highly complex proof statements like those that the Department of Defense (DoD) may wish to employ, novel and more efficient approaches are needed.
The Securing Information for Encrypted Verification and Evaluation (SIEVE) program seeks to advance the state of the art in ZK proofs to enable complex, DoD-relevant applications. SIEVE will use ZK proofs to enable the verification of capabilities relevant to the DoD without revealing the sensitive details associated with those capabilities. SIEVE will aim to accomplish this goal by dramatically increasing the expressivity of problem statements for which ZK proofs can be constructed. SIEVE will also focus on increasing the efficiency of ZK proof technology to enable large, complex proof statements (e.g., billions of gates or more, where the statement natively consists of probabilistic, indeterminate-branching conditions). SIEVE will demonstrate the feasibility of encoding complex, DoD-relevant statements into intermediate representations (IRs) that can then be used to create efficient ZK proofs for those statements.
Additionally, in order to ensure the relevance of ZK proofs for the foreseeable future, including the case where a cryptographically-relevant quantum computer were to exist, SIEVE will focus on substantially decreasing the asymptotic complexity of post-quantum ZK proof techniques, specifically ZK proofs that 1) rely on post-quantum hardness assumptions for their security and/or 2) reason about statements of relevance to post-quantum cryptography.