International Association
for Cryptologic Research

Common randomness is an essential resource in many applications. However, a celebrated result of Cleve (STOC 86) rules out the possibility of tossing a fair coin from scratch in the presence of a dishonest majority. A second-best alternative is a Coin Tossing Extension (CTE) protocol, which uses an "online" oracle that produces a small number of common random bits to generate a large number of common random-looking bits.This work initiates the systematic study of fully-secure CTE, which guarantees output even in the presence of malicious behavior. A fully-secure two-party statistical CTE protocol with black-box simulation was implicit in Hofheinz et al. (Eurocrypt 06), but its round complexity is nearly linear in its output length. The problem of constant-round CTE with superlogarithmic stretch remained open. We prove that any statistical CTE with full black-box security and superlogarithmic stretch must have superconstant rounds. To circumvent this impossibility we investigate fully-secure computational CTE, and prove that with N parties and any polynomial stretch: • One round suffices for CTE under subexponential LWE, even with Universally Composable security against adaptive corruptions. • One-round CTE is implied by DDH or the hidden subgroup assumption in class groups, with a short, reusable Uniform Random String, and by both DCR and QR, with a reusable Structured Reference String. • One-way functions imply CTE with O(N) rounds, and thus constant-round CTE for any constant number of parties. Such results were not known even in the two-party setting with standalone, static security. Furthermore, we extend one-round CTE to sample from any efficient distribution, via strong assumptions that include indistinguishability obfuscation. Our one-round CTE protocols can be interpreted as explainable variants of classical randomness extractors, wherein a (short) seed and a source instance can be efficiently reverse-sampled given a random output. Such explainable extractors may be of independent interest.

This work introduces homomorphic secret sharing (HSS) with succinct share size. In HSS, private inputs are shared between parties, who can then homomorphically evaluate a function on their shares, obtaining a share of the function output. In succinct HSS, a portion of the inputs can be distributed using shares whose size is sublinear in the number of such inputs. The parties can then locally evaluate a function f on the shares, with the restriction that f must be linear in the succinctly shared inputs. We construct succinct, two-party HSS for branching programs, based on either the decisional composite residuosity assumption, a DDH-like assumption in class groups, or learning with errors with a superpolynomial modulus-to-noise ratio. We then give several applications of succinct HSS, which were only previously known using fully homomorphic encryption, or stronger tools: 1. Succinct vector oblivious linear evaluation (VOLE): Two parties can obtain secret shares of a long, arbitrary vector x, multiplied by a scalar ∆, with communication sublinear in the size of the vector. 2. Batch, multi-party distributed point functions: A protocol for distributing a batch of secret, random point functions among N parties, for any polynomial N, with communication sublinear in the number of DPFs. 3. Sublinear MPC for any number of parties: Two new constructions of MPC with sublinear communication complexity, with N parties for any polynomial N: (1) For general layered Boolean circuits of size s, with communication O(N s/log log s), and (2) For layered, sufficiently wide Boolean circuits, with communication O(N s/log s).

A distributed sampler is a way for several mutually distrusting parties to non-interactively generate a common reference string (CRS) that all parties trust. Previous work constructs distributed samplers in the random oracle model, or in the standard model with very limited security guarantees. This is no accident, as standard model distributed samplers with full security were shown impossible. In this work, we provide new definitions for distributed samplers which we show achieve meaningful security guarantees in the standard model. In particular, our notion implies that the hardness of a wide range of security games is preserved when the CRS is replaced with a distributed sampler. We also show how to realize our notion of distributed samplers. A core technical tool enabling our construction is a new notion of single-message zero knowledge.

We study the following broad question about cryptographic primitives: is it possible to achieve security against arbitrary poly(n)-size adversaries with O(log n)-size messages? It is common knowledge that the answer is “no” unless information-theoretic security is possible. In this work, we revisit this question by considering the setting of cryptography with public information and computational security. We obtain the following main results, assuming variants of well-studied intractability assumptions: 1. A private simultaneous messages (PSM) protocol for every f : [n] × [n] → {0, 1} with (1 + eps) log n-bit messages, beating the known lower bound on information-theoretic PSM. We apply this towards non-interactive secure 3-party computation with similar message size in the preprocessing model, improving over previous 2-round protocols. 2. A secret-sharing scheme for any “forbidden-graph” access structure on n nodes with O(log n) share size. 3. On the negative side, we show that computational threshold secret-sharing schemes with public information require share size Ω(log log n). For arbitrary access structures, we show that computational security does not help with 1-bit shares. The above positive results guarantee that any adversary of size n^{o(log n)} achieves an n^{−Ω(1)} distinguishing advantage. We show how to make the advantage negligible by slightly increasing the asymptotic message size, still improving over all known constructions. The security of our constructions is based on the conjectured hardness of variants of the planted clique problem, which was extensively studied in the algorithms, statistical inference, and complexity-theory communities. Our work provides the first applications of such assumptions to improving the efficiency of mainstream cryptographic primitives, gives evidence for the necessity of such assumptions, and gives rise to new questions in this domain that may be of independent interest.

The SPDZ protocol for multi-party computation relies on a correlated randomness setup consisting of authenticated, multiplication triples. A recent line of work by Boyle et al. (Crypto 2019, Crypto 2020) has investigated the possibility of producing this correlated randomness in a \emph{silent preprocessing} phase, which involves a ``small'' setup protocol with less communication than the total size of the triples being produced. These works do this using a tool called a \emph{pseudorandom correlation generator} (PCG), which allows a large batch of correlated randomness to be compressed into a set of smaller, correlated seeds. However, existing methods for compressing SPDZ triples only apply to the 2-party setting. In this work, we construct a PCG for producing SPDZ triples over large prime fields in the multi-party setting. The security of our PCG is based on the ring-LPN assumption over fields, similar to the work of Boyle et al. (Crypto 2020) in the 2-party setting. We also present a corresponding, actively secure setup protocol, which can be used to generate the PCG seeds and instantiate SPDZ with a silent preprocessing phase. As a building block, which may be of independent interest, we construct a new type of 3-party distributed point function supporting outputs over arbitrary groups (including large prime order), as well as an efficient protocol for setting up our DPF keys with active security.

Structured random strings (SRSs) and correlated randomness are important for many cryptographic protocols. In settings where interaction is expensive, it is desirable to obtain such randomness in as few rounds of communication as possible; ideally, simply by exchanging one reusable round of messages which can be considered public keys. In this paper, we describe how to generate any SRS or correlated randomness in such a single round of communication, using, among other things, indistinguishable obfuscation. We introduce what we call a distributed sampler, which enables n parties to sample a single public value (SRS) from any distribution. We construct a semi-malicious distributed sampler in the plain model, and use it to build a semi-malicious public- key PCF (Boyle et al., FOCS 2020) in the plain model. A public-key PCF can be thought of as a distributed correlation sampler; instead of producing a public SRS, it gives each party a private random value (where the values satisfy some correlation). We introduce a general technique called an anti-rusher which compiles any one-round protocol with semi-malicious security without inputs to a similar one-round protocol with active security by making use of a programmable random oracle. This gets us actively secure distributed samplers and public-key PCFs in the random oracle model. Finally, we explore some tradeoffs. Our first PCF construction is limited to reverse-sampleable correlations (where the random outputs of honest parties must be simulatable given the random outputs of corrupt parties); we additionally show a different construction without this limitation, but which does not allow parties to hold secret parameters of the correlation. We also describe how to avoid the use of a random oracle at the cost of relying on sub-exponentially secure indistinguishability obfuscation.

Recently, number-theoretic assumptions including DDH, DCR and QR have been used to build powerful tools for secure computation, in the form of homomorphic secret-sharing (HSS), which leads to secure two-party computation protocols with succinct communication, and pseudorandom correlation functions (PCFs), which allow non-interactive generation of a large quantity of correlated randomness. In this work, we present a group-theoretic framework for these classes of constructions, which unifies their approach to computing distributed discrete logarithms in various groups. We cast existing constructions in our framework, and also present new constructions, including one based on class groups of imaginary quadratic fields. This leads to the first construction of two-party homomorphic secret sharing for branching programs from class group assumptions. Using our framework, we also obtain pseudorandom correlation functions for generating oblivious transfer and vector-OLE correlations from number-theoretic assumptions. These have a trustless, public-key setup when instantiating our framework using class groups. Previously, such constructions either needed a trusted setup in the form of an RSA modulus with unknown factorisation, or relied on multi-key fully homomorphic encryption from the learning with errors assumption. We also show how to upgrade our constructions to achieve active security using appropriate zero-knowledge proofs. In the random oracle model, this leads to a one-round, actively secure protocol for setting up the PCF, as well as a 3-round, actively secure HSS-based protocol for secure two-party computation of branching programs with succinct communication.