This document summarizes an improved protocol for secure two-party computation using additive homomorphic encryption. The proposed protocol allows a server to compute an encrypted function f(m) on an encrypted input Enc(m), without learning the plaintext value m, with the help of a client who holds the decryption key. The improvement reduces the protocol to 1 round of communication, compared to 2 rounds previously, and reduces the communication size by a factor of 40-130x. It achieves this by having the client use zero-knowledge proofs to convince the server that it correctly generated the output encryption Enc'(f(m)) from the input encryption Enc(m), without revealing any plaintext values.
zk-SNARKs are zero-knowledge succinct non-interactive arguments of knowledge that allow a prover to convince a verifier of a statement without revealing details. They work by converting a function and its inputs/outputs into a quadratic arithmetic program (QAP) represented as polynomials. This allows a verifier to efficiently check a proof generated by the prover using techniques like Lagrange interpolation and pairings on elliptic curves to ensure the polynomials satisfy the QAP without directly evaluating the function. The setup requires a "trusted setup" but then allows very efficient verification.
The document discusses the SATYSFI Conf 2021 conference which will take place on June 26, 2021. It provides details on recent updates to the SATYSFI typesetting system including the addition of linear-transform-graphics, improvements to page breaking for multicolumn content, and adding debugging information for overfull/underfull boxes. Version 0.0.6, 0.0.7, and planned future updates are summarized. The document also discusses using domain specific languages for describing typesetting definitions and structures and provides examples using amidakuji diagrams.
This document summarizes an improved protocol for secure two-party computation using additive homomorphic encryption. The proposed protocol allows a server to compute an encrypted function f(m) on an encrypted input Enc(m), without learning the plaintext value m, with the help of a client who holds the decryption key. The improvement reduces the protocol to 1 round of communication, compared to 2 rounds previously, and reduces the communication size by a factor of 40-130x. It achieves this by having the client use zero-knowledge proofs to convince the server that it correctly generated the output encryption Enc'(f(m)) from the input encryption Enc(m), without revealing any plaintext values.
zk-SNARKs are zero-knowledge succinct non-interactive arguments of knowledge that allow a prover to convince a verifier of a statement without revealing details. They work by converting a function and its inputs/outputs into a quadratic arithmetic program (QAP) represented as polynomials. This allows a verifier to efficiently check a proof generated by the prover using techniques like Lagrange interpolation and pairings on elliptic curves to ensure the polynomials satisfy the QAP without directly evaluating the function. The setup requires a "trusted setup" but then allows very efficient verification.
The document discusses the SATYSFI Conf 2021 conference which will take place on June 26, 2021. It provides details on recent updates to the SATYSFI typesetting system including the addition of linear-transform-graphics, improvements to page breaking for multicolumn content, and adding debugging information for overfull/underfull boxes. Version 0.0.6, 0.0.7, and planned future updates are summarized. The document also discusses using domain specific languages for describing typesetting definitions and structures and provides examples using amidakuji diagrams.