International Association
for Cryptologic Research

Lattice-based cryptography, the study of cryptographic primitives whose security is based on the hardness of so-called lattice problems, has taken center stage in cryptographic research in recent years. It potentially offers favorable security features, even against quantum algorithms. One of the main obstacles for wide adoption of this type of cryptography is its unsatisfactory efficiency. To address this point, efficient lattice-based cryptography usually relies on the intractability of problems on lattices with additional algebraic structure (such as so-called ideal-lattices or module-lattices). It is an important open question to evaluate the hardness of such lattice problems, and their relation to the hardness of problems on unstructured lattices. It is a known fact that an unstructured lattice, which is simply an additive discrete group in Euclidean space, can be cast as an ideal-lattice in some \emph{order} of a number field (and thus, in a rather trivial sense, that ideals in orders are as general as unstructured lattices). However, it is not known whether this connection can be used to imply useful hardness results for structured lattices, or alternatively new algorithmic techniques for unstructured lattices. In this work we establish a gradient of hardness for the Order-LWE problem (a generalization of the well known Ring-LWE problem), as it varies over orders in a number field. Furthermore, we show that, in every number field, there are certain orders such that the corresponding Order-LWE problem is at least as hard as the (unstructured) LWE problem. So in general one should not hope to solve (any) Order-LWE more efficiently than LWE. However, we show that this connection holds in orders that are very ``skewed'' and hence, perhaps, irrelevant for improving efficiency in cryptographic applications. We further improve the hardness result for Order-LWE, to include \textit{all} ideal lattices, closing a gap left in prior work. This establishes a direct connection between problems on unstructured lattices and the structured problem of Order-LWE.

This work addresses the long quest for proving full (adaptive) security for attribute-based encryption (ABE). We show that in order to prove full security in a black-box manner, the scheme must be "irregular" in the sense that it is impossible to "validate" secret keys to ascertain consistent decryption of ciphertexts. This extends a result of Lewko and Waters (Eurocrypt 2014) that was only applicable to straight-line proofs (without rewinding). Our work, therefore, establishes that it is impossible to circumvent the irregularity property using creative proof techniques, so long as the adversary is used in a black-box manner. As a consequence, our work provides an explanation as to why some lattice-based ABE schemes cannot be proven fully secure, even though no known adaptive attacks exist.

We explore a very simple distribution of unitaries: random (binary) phase -- Hadamard -- random (binary) phase -- random computational-basis permutation. We show that this distribution is statistically indistinguishable from random Haar unitaries for any polynomial set of orthogonal input states (in any basis) with polynomial multiplicity. This shows that even though real-valued unitaries cannot be completely pseudorandom (Haug, Bharti, Koh, arXiv:2306.11677), we can still obtain some pseudorandom properties without giving up on the simplicity of a real-valued unitary. Our analysis shows that an even simpler construction: applying a random (binary) phase followed by a random computational-basis permutation, would suffice, assuming that the input is orthogonal and \emph{flat} (that is, has high min-entropy when measured in the computational basis). Using quantum-secure one-way functions (which imply quantum-secure pseudorandom functions and permutations), we obtain an efficient cryptographic instantiation of the above.

A test of quantumness is a protocol that allows a classical verifier to certify (only) that a prover is not classical. We show that tests of quantumness that follow a certain template, which captures recent proposals such as [KCVY21,KLVY22], in fact can do much more. Namely, the same protocols can be used for certifying a qubit, a building-block that stands at the heart of applications such as certifiable randomness and classical delegation of quantum computation. Certifying qubits was previously only known to be possible based on the hardness of the Learning with Errors problem and the use of adaptive hardcore bits [BCM+21]. Our framework allows certification of qubits based only on the existence of post-quantum trapdoor claw-free functions, or on quantum fully homomorphic encryption. These can be instantiated, for example, from Ring Learning with Errors. This has the potential to improve the efficiency of qubit certification and derived functionalities. On the technical side, we show that the quantum soundness of any such protocol can be reduced to proving a bound on a simple algorithmic task: informally, answering “two challenges simultaneously” in the protocol. Our reduction formalizes the intuition that these protocols demonstrate quantumness by leveraging the impossibility of rewinding a general quantum prover. This allows us to prove tight bounds on the quantum soundness of [KCVY21] and [KLVY22], showing that no quantum polynomial-time prover can succeed with probability larger than cos2 π8 ≈ 0.853. Previously, only an upper bound on the success probability of classical provers, and a lower bound on the success probability of quantum provers, were known. We then extend this proof of quantum soundness to show that provers that approach the quantum soundness bound must perform almost anti-commuting measurements. This certifies that the prover holds a qubit.

We propose to study equivalence relations between phenomena in high-energy physics and the existence of standard cryptographic primitives, and show the first example where such an equivalence holds. A small number of prior works showed that high-energy phenomena can be explained by cryptographic hardness. Examples include using the existence of one-way functions to explain the hardness of decoding black-hole Hawking radiation (Harlow and Hayden 2013, Aaronson 2016), and using pseudorandom quantum states to explain the hardness of computing AdS/CFT dictionary (Bouland, Fefferman and Vazirani, 2020). In this work we show, for the former example of black-hole radiation decoding, that it also implies the existence of secure quantum cryptography. In fact, we show an existential equivalence between the hardness of black-hole radiation decoding and a variety of cryptographic primitives, including bit-commitment schemes and oblivious transfer protocols (using quantum communication). This can be viewed (with proper disclaimers, as we discuss) as providing a physical justification for the existence of secure cryptography. We conjecture that such connections may be found in other high-energy physics phenomena.

We construct a succinct non-interactive argument ($\mathsf{SNARG}$) for the class of monotone policy batch $\mathsf{NP}$ languages, under the Learning with Errors ($\mathsf{LWE}$) assumption. This class is a subclass of $\mathsf{NP}$ that is associated with a monotone function~$f:\{0,1\}^k\rightarrow\{0,1\}$ and an $\mathsf{NP}$ language $\mathcal{L}$, and contains instances $(x_1,\ldots,x_k)$ such the $f(b_1,\ldots,b_k)=1$ where $b_j=1$ if and only if $x_j\in \mathcal{L}$. Our $\mathsf{SNARG}$s are arguments of knowledge in the non-adaptive setting, and satisfy a new notion of somewhere extractability against adaptive adversaries. This is the first $\mathsf{SNARG}$ under standard hardness assumptions for a sub-class of $\mathsf{NP}$ that is not known to have a (computational) non-signaling $\mathsf{PCP}$ with parameters compatible with the standard framework for constructing $\mathsf{SNARG}$s dating back to [Kalai-Raz-Rothblum, STOC '13]. Indeed, our approach necessarily departs from this framework. Our construction combines existing quasi-arguments for $\mathsf{NP}$ (based on batch arguments for $\mathsf{NP}$) with a new type of cryptographic encoding of the instance and a new analysis going from local to global soundness. The main novel ingredient used in our encoding is a {\em predicate-extractable hash} ($\mathsf{PEH}$) family, which is a primitive that generalizes the notion of a somewhere extractable hash. Whereas a somewhere extractable hash allows to extract a single input coordinate, our $\mathsf{PEH}$ extracts a {\em global} property of the input. We view this primitive to be of independent interest, and believe that it will find other applications.

We propose a new approach to construct general-purpose indistinguishability obfuscation (iO). Our construction is obtained via a new intermediate primitive that we call split fully homomorphic encryption (split FHE), which we show to be sufficient for constructing iO. Specifically, split FHE is FHE where decryption takes the following two-step syntactic form: (i) a secret decryption step that uses the secret key and produces a hint which is (asymptotically) shorter than the length of the encrypted message, and (ii) a public decryption step that only requires the ciphertext and the previously generated hint (and not the entire secret key) and recovers the encrypted message. In terms of security, the hints for a set of ciphertexts should not allow one to violate semantic security for any other ciphertexts. Next, we show a generic candidate construction of split FHE based on three building blocks: (i) A standard FHE scheme with linear decrypt-and-multiply (which can be instantiated with essentially all LWE-based constructions), (ii) a linearly homomorphic encryption scheme with short decryption hints (such as the Damgård-Jurik encryption scheme, based on the DCR problem), and (iii) a cryptographic hash function (which can be based on a variety of standard assumptions). Our approach is heuristic in the sense that our construction is not provably secure and makes implicit assumptions about the interplay between these underlying primitives. We show evidence that this construction is secure by providing an argument in an appropriately defined oracle model. We view our construction as a big departure from the state-of-the-art constructions, and it is in fact quite simple.

Two fundamental properties of quantum states that quantum information theory explores are pseudorandomness and provability of destruction. We introduce the notion of quantum pseudorandom states with proofs of destruction (PRSPD) that combines both these properties. Like standard pseudorandom states (PRS), these are efficiently generated quantum states that are indistinguishable from random, but they can also be measured to create a classical string. This string is verifiable (given the secret key) and certifies that the state has been destructed. We show that, similarly to PRS, PRSPD can be constructed from any post-quantum one-way function. As far as the authors are aware, this is the first construction of a family of states that satisfies both pseudorandomness and provability of destruction. We show that many cryptographic applications that were shown based on PRS variants using quantum communication can be based on (variants of) PRSPD using only classical communication. This includes symmetric encryption, message authentication, one-time signatures, commitments, and classically verifiable private quantum coins.

We show that it is possible to perform $n$ independent copies of $1$-out-of-$2$ oblivious transfer in two messages, where the communication complexity of the receiver and sender (each) is $n(1+o(1))$ for sufficiently large $n$. Note that this matches the information-theoretic lower bound. Prior to this work, this was only achievable by using the heavy machinery of rate-$1$ fully homomorphic encryption (Rate-$1$ FHE, Brakerski et al., TCC 2019). To achieve rate-$1$ both on the receiver's and sender's end, we use the LPN assumption, with slightly sub-constant noise rate $1/m^{\epsilon}$ for any $\epsilon>0$ together with either the DDH, QR or LWE assumptions. In terms of efficiency, our protocols only rely on linear homomorphism, as opposed to the FHE-based solution which inherently requires an expensive ``bootstrapping'' operation. We believe that in terms of efficiency we compare favorably to existing batch-OT protocols, while achieving superior communication complexity. We show similar results for Oblivious Linear Evaluation (OLE). For our DDH-based solution we develop a new technique that may be of independent interest. We show that it is possible to ``emulate'' the binary group $\bbZ_2$ (or any other small-order group) inside a prime-order group $\bbZ_p$ \emph{in a function-private manner}. That is, $\bbZ_2$ operations are mapped to $\bbZ_p$ operations such that the outcome of the latter do not reveal additional information beyond the $\bbZ_2$ outcome. Our encoding technique uses the discrete Gaussian distribution, which to our knowledge was not done before in the context of DDH.

Is it possible to convert classical reductions into post-quantum ones? It is customary to argue that while this is problematic in the interactive setting, non-interactive reductions do carry over. However, when considering quantum auxiliary input, this conversion results in a *non-constructive* post-quantum reduction that requires duplicating the quantum auxiliary input, which is in general inefficient or even impossible. This violates the win-win premise of provable cryptography: an attack against a cryptographic primitive should lead to an algorithmic advantage. We initiate the study of constructive quantum reductions and present positive and negative results for converting large classes of classical reductions to the post-quantum setting in a constructive manner. We show that any non-interactive non-adaptive reduction from assumptions with a polynomial solution space (such as decision assumptions) can be made post-quantum constructive. In contrast, assumptions with super-polynomial solution space (such as general search assumptions) cannot be generally converted. Along the way, we make several additional contributions: 1. We put forth a framework for reductions (or general interaction) with *stateful* solvers for a computational problem, that may change their internal state between consecutive calls. We show that such solvers can still be utilized. This framework and our results are meaningful even in the classical setting. 2. A consequence of our negative result is that quantum auxiliary input that is useful against a problem with a super-polynomial solution space cannot be generically ``restored'' post-measurement. This shows that the novel rewinding technique of Chiesa et al.\ (FOCS 2021) is tight in the sense that it cannot be extended beyond a polynomial measurement space.

Virtual black-box obfuscation is a strong cryptographic primitive: it encrypts a circuit while maintaining its full input/output functionality. A remarkable result by Barak et al. (Crypto 2001) shows that a general obfuscator that obfuscates classical circuits into classical circuits cannot exist. A promising direction that circumvents this impossibility result is to obfuscate classical circuits into quantum states, which would potentially be better capable of hiding information about the obfuscated circuit. We show that, under the assumption that Learning With Errors (LWE) is hard for quantum computers, this quantum variant of virtual black-box obfuscation of classical circuits is generally impossible. On the way, we show that under the presence of dependent classical auxiliary input, even the small class of classical point functions cannot be quantum virtual black-box obfuscated.

In classical commitments, statistical binding means that for almost any commitment transcript there is at most one possible opening. While quantum commitments (for classical messages) sometimes have benefits over their classical counterparts (e.g. in terms of assumptions), they provide a weaker notion of binding. Essentially that the sender cannot open a given commitment to a random value with probability noticeably greater than 1/2. We introduce a notion of classical binding for quantum commitments which provides guarantees analogous to the classical case. In our notion, the receiver performs a (partial) measurement of the quantum commitment string, and the outcome of this measurement determines a single value that the sender may open. We expect that our notion can replace classical commitments in various settings, leaving the security proof essentially unchanged. As an example we show a soundness proof for the GMW zero-knowledge proof system. We construct a non-interactive quantum commitment scheme which is classically statistically-binding and has a classical opening, based on the existence of any post-quantum one-way function. Prior candidates had inherently quantum openings and were not classically binding. In contrast, we show that it is impossible to achieve classical binding for statistically hiding commitments, regardless of assumption or round complexity. Our scheme is simply Naor's commitment scheme (which classically requires a common random string, CRS), but executed in superposition over all possible values of the CRS, and repeated several times. We hope that this technique for using quantum communication to remove a CRS may find other uses.

We present a candidate obfuscator based on composite-order graded encoding schemes (GES), which are a generalization of multilinear maps. Our obfuscator operates on circuits directly without converting them into formulas or branching programs as was done in previous solutions. As a result, the time and size complexity of the obfuscated program, measured by the number of GES elements, is directly proportional to the circuit complexity of the program being obfuscated. This improves upon previous constructions whose complexity was related to the formula or branching program size. Known instantiations of Graded Encoding Schemes allow us to obfuscate circuit classes of polynomial degree, which include for example families of circuits of logarithmic depth. We prove that our obfuscator is secure against a class of generic algebraic attacks, formulated by a generic graded encoding model. We further consider a more robust model which provides more power to the adversary and extend our results to this setting as well. As a secondary contribution, we define a new simple notion of algebraic security (which was implicit in previous works) and show that it captures standard security relative to an ideal GES oracle.

The hardness of the Learning with Errors (LWE) problem is by now a cornerstone of the cryptographic landscape, allowing to con- struct cryptographic schemes with properties unknown under other as- sumptions, and being conjectured to be resilient to quantum attacks. LWE is essentially the task of solving a noisy system of random linear equations over uniformly random secret variables (“the LWE secret”), evaluated modulo some integer. In applications the secret variables usu- ally correspond to the secret key of the cryptographic scheme. It is therefore of great importance to understand what happens when the secret variables are not sampled uniformly (but still have some entropy). This is relevant for settings where an adversary manages to obtain partial information on the secret (a.k.a key leakage), for various theoretical ap- plications, and also for practical use where for efficiency or convenience it is easier to sample the secret from some non-uniform distribution. This so called “Entropic LWE” problem has been studied in a number of works, starting with Goldwasser et al. (ICS 2010). However, so far it was only known how to prove the hardness of Entropic LWE for secret distributions supported inside a ball of small radius. In this work we resolve the hardness of Entropic LWE with arbitrary long secrets, in the following sense. We show an entropy bound that guarantees the security of arbitrary Entropic LWE. This bound is higher than what is required in the ball-bounded setting, but we show that this is essentially tight. Tightness is shown unconditionally for highly-composite moduli, and using black-box impossibility for arbitrary moduli. Technically, we show that the entropic hardness of LWE relies on a sim- ple to describe lossiness property of the distribution of secrets itself. This is simply the probability of recovering a random sample from this distri- bution s, given s + e, where e is Gaussian noise (i.e. the quality of the distribution of secrets as an error correcting code for Gaussian noise). We hope that this characterization will make it easier to derive entropic LWE results more easily in the future. We also use our techniques to show new results for the ball-bounded setting, essentially showing that under a strong enough assumption even polylogarithmic entropy suffices.

We propose a new approach to construct general-purpose indistinguishability obfuscation (iO). Our construction is obtained via a new intermediate primitive that we call split fully-homomorphic encryption (split FHE), which we show to be sufficient for constructing iO. Specifically, split FHE is FHE where decryption takes the following two-step syntactic form: (i) A secret decryption step uses the secret key and produces a hint which is (asymptotically) shorter than the length of the encrypted message, and (ii) a public decryption step that only requires the ciphertext and the previously generated hint (and not the entire secret key), and recovers the encrypted message. In terms of security, the hints for a set of ciphertexts should not allow one to violate semantic security for any other ciphertexts. Next, we show a generic candidate construction of split FHE based on three building blocks: (i) A standard FHE scheme with linear decrypt-and-multiply (which can be instantiated with essentially all LWE-based constructions), (ii) a linearly homomorphic encryption scheme with short decryption hints (such as the Damgard-Jurik encryption scheme, based on the DCR problem), and (iii) a cryptographic hash function (which can be based on a variety of standard assumptions). Our approach is heuristic in the sense that our construction is not provably secure and makes implicit assumptions about the interplay between these underlying primitives. We show evidence that this construction is secure by providing an argument in an appropriately defined oracle model. We view our construction as a big departure from the state-of-the-art constructions, and it is in fact quite simple.

Consider an access policy for some resource which only allows access to users of the system who own a certain set of attributes. Specifically, we consider the case where such an access structure is defined by some monotone function $$f:{0,1}^N ightarrow {0,1}$$ , belonging to some class of function $$F$$ (e.g. conjunctions, space bounded computation), where N is the number of possible attributes. In this work we show that any succinct single-round delegation scheme for the function class $$F$$ can be converted into a succinct single-round private access control protocol. That is, a verifier can be convinced that an approved user (i.e. one which holds an approved set of attributes) is accessing the system, without learning any additional information about the user or the set of attributes. As a main tool of independent interest, we show that assuming a quasi-polynomially secure two-message oblivious transfer scheme with statistical sender privacy (which can be based on quasi-polynomial hardness of the DDH, QR, DCR or LWE assumptions), we can convert any single-round protocol into a witness indistinguishable one, with similar communication complexity.

Efficiently sampling a quantum state that is hard to distinguish from a truly random quantum state is an elementary task in quantum information theory that has both computational and physical uses. This is often referred to as pseudorandom (quantum) state generator, or PRS generator for short. In existing constructions of PRS generators, security scales with the number of qubits in the states, i.e.\ the (statistical) security parameter for an $n$-qubit PRS is roughly $n$. Perhaps counter-intuitively, $n$-qubit PRS are not known to imply $k$-qubit PRS even for $k<n$. Therefore the question of \emph{scalability} for PRS was thus far open: is it possible to construct $n$-qubit PRS generators with security parameter $\secp$ for all $n, \secp$. Indeed, we believe that PRS with tiny (even constant) $n$ and large $\secp$ can be quite useful. We resolve the problem in this work, showing that any quantum-secure one-way function implies scalable PRS. We follow the paradigm of first showing a \emph{statistically} secure construction when given oracle access to a random function, and then replacing the random function with a quantum-secure (classical) pseudorandom function to achieve computational security. However, our methods deviate significantly from prior works since scalable pseudorandom states require randomizing the amplitudes of the quantum state, and not just the phase as in all prior works. We show how to achieve this using Gaussian sampling.

We present new Non-Interactive Zero-Knowledge argument systems (NIZK), based on standard assumptions that were previously not known to imply it. In particular, we rely on the hardness of both the learning parity with noise (LPN) assumption, and the existence of trapdoor hash functions (TDH, defined by Döttling et al., Crypto 2019). TDH can be based on a number of standard assumptions, including DDH, QR, DCR, and LWE. We rely on the Correlation Intractability (CI) framework for converting \Sigma-protocols into NIZK, but deviate from prior works in considering CI for searchable relations where the search function has a probabilistic representation by a simple function class (linear or constant degree in our instantiations). Namely, there is a distribution over simple functions that computes each output bit of the search function with all but small (constant) probability. We present a new tool for proving CI for such function classes via a notion that we call Approximate-Correlation Intractability. This notion requires that CI holds even against approximations of a given function class. We show that approximate-correlation intractability for just constant degree functions suffices if the underlying \Sigma-protocol is implemented using an extractable commitment scheme with approximately low-degree extraction, and that such a commitment scheme can be constructed based on LPN. We then show how to construct approximate CI hash functions for this class from any suitable rate-1 TDH (with an enhanced correctness property that is satisfied by all existing constructions).

Non-committing encryption (NCE) is a type of public key encryption which comes with the ability to equivocate ciphertexts to encryptions of arbitrary messages, i.e., it allows one to find coins for key generation and encryption which ``explain'' a given ciphertext as an encryption of any message. NCE is the cornerstone to construct adaptively secure multiparty computation [Canetti et al. STOC'96] and can be seen as the quintessential notion of security for public key encryption to realize ideal communication channels. A large body of literature investigates what is the best message-to-ciphertext ratio (i.e., the rate) that one can hope to achieve for NCE. In this work we propose a near complete resolution to this question and we show how to construct NCE with constant rate in the plain model from a variety of assumptions, such as the hardness of the learning with errors (LWE), the decisional Diffie-Hellman (DDH), or the quadratic residuosity (QR) problem. Prior to our work, constructing NCE with constant rate required a trusted setup and indistinguishability obfuscation [Canetti et al. ASIACRYPT'17].

The hardness of the Ring Learning with Errors problem (RLWE) is a central building block for efficiency-oriented lattice-based cryptography. Many applications use an ``entropic'' variant of the problem where the so-called ``secret'' is not distributed uniformly as prescribed but instead comes from some distribution with sufficient min-entropy. However, the hardness of the entropic variant has not been substantiated thus far. For standard LWE (not over rings) entropic results are known, using a ``lossiness approach'' but it was not known how to adapt this approach to the ring setting. In this work we present the first such results, where entropic security is established either under RLWE or under the Decisional Small Polynomial Ratio (DSPR) assumption which is a mild variant of the NTRU assumption. In the context of general entropic distributions, our results in the ring setting essentially match the known lower bounds (Bolboceanu et al., Asiacrypt 2019; Brakerski and Döttling, Eurocrypt 2020).

We present a novel tree-based technique that can convert any designated-prover NIZK proof system (DP-NIZK) which maintains zero-knowledge only for single statement, into one that allows to prove an unlimited number of statements in ZK, while maintaining all parameters succinct. Our transformation requires leveled fully-homomorphic encryption. We note that single-statement DP-NIZK can be constructed from any one-way function. We also observe a two-way derivation between DP-NIZK and attribute-based signatures (ABS), and as a result derive now constructions of ABS and homomorphic signatures (HS). Our construction improves upon the prior construction of lattice-based DP-NIZK by Kim and Wu (Crypto 2018) since we only require leveled FHE as opposed to HS (which also translates to improved LWE parameters when instantiated). Alternatively, the recent construction of NIZK without preprocessing from either circular-secure FHE (Canetti et al., STOC 2019) or polynomial Learning with Errors (Peikert and Shiehian, Crypto 2019) could be used to obtain a similar final statement. Nevertheless, we note that our statement is formally incomparable to these works (since leveled FHE is not known to imply circular secure FHE or the hardness of LWE). We view this as evidence for the potential in our technique, which we hope can find additional applications in future works.

We show, via a non-interactive reduction, that the existence of a secure multi-party computation (MPC) protocol for degree-2 functions implies the existence of a protocol with the same round complexity for general functions. Thus showing that when considering the round complexity of MPC, it is sufficient to consider very simple functions.Our completeness theorem applies in various settings: information theoretic and computational, fully malicious and malicious with various types of aborts. In fact, we give a master theorem from which all individual settings follow as direct corollaries. Our basic transformation does not require any additional assumptions and incurs communication and computation blow-up which is polynomial in the number of players and in $$S,2^D$$S,2D, where S, D are the circuit size and depth of the function to be computed. Using one-way functions as an additional assumption, the exponential dependence on the depth can be removed.As a consequence, we are able to push the envelope on the state of the art in various settings of MPC, including the following cases. 3-round perfectly-secure protocol (with guaranteed output delivery) against an active adversary that corrupts less than 1/4 of the parties.2-round statistically-secure protocol that achieves security with “selective abort” against an active adversary that corrupts less than half of the parties.Assuming one-way functions, 2-round computationally-secure protocol that achieves security with (standard) abort against an active adversary that corrupts less than half of the parties. This gives a new and conceptually simpler proof to the recent result of Ananth et al. (Crypto 2018). Technically, our non-interactive reduction draws from the encoding method of Applebaum, Brakerski and Tsabary (TCC 2018). We extend these methods to ones that can be meaningfully analyzed even in the presence of malicious adversaries.

In (single-server) Private Information Retrieval (PIR), a server holds a large database $${\mathtt {DB}}$$ of size n, and a client holds an index $$i \in [n]$$ and wishes to retrieve $${\mathtt {DB}}[i]$$ without revealing i to the server. It is well known that information theoretic privacy even against an “honest but curious” server requires $$\varOmega (n)$$ communication complexity. This is true even if quantum communication is allowed and is due to the ability of such an adversarial server to execute the protocol on a superposition of databases instead of on a specific database (“input purification attack”).Nevertheless, there have been some proposals of protocols that achieve sub-linear communication and appear to provide some notion of privacy. Most notably, a protocol due to Le Gall (ToC 2012) with communication complexity $$O(\sqrt{n})$$ , and a protocol by Kerenidis et al. (QIC 2016) with communication complexity $$O(\log (n))$$ , and O(n) shared entanglement.We show that, in a sense, input purification is the only potent adversarial strategy, and protocols such as the two protocols above are secure in a restricted variant of the quantum honest but curious (a.k.a specious) model. More explicitly, we propose a restricted privacy notion called anchored privacy, where the adversary is forced to execute on a classical database (i.e. the execution is anchored to a classical database). We show that for measurement-free protocols, anchored security against honest adversarial servers implies anchored privacy even against specious adversaries.Finally, we prove that even with (unlimited) pre-shared entanglement it is impossible to achieve security in the standard specious model with sub-linear communication, thus further substantiating the necessity of our relaxation. This lower bound may be of independent interest (in particular recalling that PIR is a special case of Fully Homomorphic Encryption).

We present a worst case decoding problem whose hardness reduces to that of solving the Learning Parity with Noise (LPN) problem, in some parameter regime. Prior to this work, no worst case hardness result was known for LPN (as opposed to syntactically similar problems such as Learning with Errors). The caveat is that this worst case problem is only mildly hard and in particular admits a quasi-polynomial time algorithm, whereas the LPN variant used in the reduction requires extremely high noise rate of $$1/2-1/\mathrm{poly}(n)$$ . Thus we can only show that “very hard” LPN is harder than some “very mildly hard” worst case problem. We note that LPN with noise $$1/2-1/\mathrm{poly}(n)$$ already implies symmetric cryptography.Specifically, we consider the (n, m, w)-nearest codeword problem ((n, m, w)-NCP) which takes as input a generating matrix for a binary linear code in m dimensions and rank n, and a target vector which is very close to the code (Hamming distance at most w), and asks to find the codeword nearest to the target vector. We show that for balanced (unbiased) codes and for relative error $$w/m \approx {\log ^2 n}/{n}$$ , (n, m, w)-NCP can be solved given oracle access to an LPN distinguisher with noise ratio $$1/2-1/\mathrm{poly}(n)$$ .Our proof relies on a smoothing lemma for codes which we show to have further implications: We show that (n, m, w)-NCP with the aforementioned parameters lies in the complexity class $$\mathrm {{Search}\hbox {-}\mathcal {BPP}}^\mathcal {SZK}$$ (i.e. reducible to a problem that has a statistical zero knowledge protocol) implying that it is unlikely to be $$\mathcal {NP}$$ -hard. We then show that the hardness of LPN with very low noise rate $$\log ^2(n)/n$$ implies the existence of collision resistant hash functions (our aforementioned result implies that in this parameter regime LPN is also in $$\mathcal {BPP}^\mathcal {SZK}$$ ).

We prove a quantum information-theoretic conjecture due to Ji, Liu and Song (CRYPTO 2018) which suggested that a uniform superposition with random binary phase is statistically indistinguishable from a Haar random state. That is, any polynomial number of copies of the aforementioned state is within exponentially small trace distance from the same number of copies of a Haar random state.As a consequence, we get a provable elementary construction of pseudorandom quantum states from post-quantum pseudorandom functions. Generating pseudorandom quantum states is desirable for physical applications as well as for computational tasks such as quantum money. We observe that replacing the pseudorandom function with a (2t)-wise independent function (either in our construction or in previous work), results in an explicit construction for quantum state t-designs for all t. In fact, we show that the circuit complexity (in terms of both circuit size and depth) of constructing t-designs is bounded by that of (2t)-wise independent functions. Explicitly, while in prior literature t-designs required linear depth (for $$t > 2$$), this observation shows that polylogarithmic depth suffices for all t.We note that our constructions yield pseudorandom states and state designs with only real-valued amplitudes, which was not previously known. Furthermore, generating these states require quantum circuit of restricted form: applying one layer of Hadamard gates, followed by a sequence of Toffoli gates. This structure may be useful for efficiency and simplicity of implementation.

We show how to combine a fully-homomorphic encryption scheme with linear decryption and a linearly-homomorphic encryption schemes to obtain constructions with new properties. Specifically, we present the following new results. (1)Rate-1 Fully-Homomorphic Encryption: We construct the first scheme with message-to-ciphertext length ratio (i.e., rate) $$1-\sigma $$ for $$\sigma = o(1)$$. Our scheme is based on the hardness of the Learning with Errors (LWE) problem and $$\sigma $$ is proportional to the noise-to-modulus ratio of the assumption. Our building block is a construction of a new high-rate linearly-homomorphic encryption.One application of this result is the first general-purpose secure function evaluation protocol in the preprocessing model where the communication complexity is within additive factor of the optimal insecure protocol.(2)Fully-Homomorphic Time-Lock Puzzles: We construct the first time-lock puzzle where one can evaluate any function over a set of puzzles without solving them, from standard assumptions. Prior work required the existence of sub-exponentially hard indistinguishability obfuscation.

We propose a generalization of the celebrated Ring Learning with Errors (RLWE) problem (Lyubashevsky, Peikert and Regev, Eurocrypt 2010, Eurocrypt 2013), wherein the ambient ring is not the ring of integers of a number field, but rather an order (a full rank subring). We show that our Order-LWE problem enjoys worst-case hardness with respect to short-vector problems in invertible-ideal lattices of the order.The definition allows us to provide a new analysis for the hardness of the abundantly used Polynomial-LWE (PLWE) problem (Stehlé et al., Asiacrypt 2009), different from the one recently proposed by Rosca, Stehlé and Wallet (Eurocrypt 2018). This suggests that Order-LWE may be used to analyze and possibly design useful relaxations of RLWE.We show that Order-LWE can naturally be harnessed to prove security for RLWE instances where the “RLWE secret” (which often corresponds to the secret-key of a cryptosystem) is not sampled uniformly as required for RLWE hardness. We start by showing worst-case hardness even if the secret is sampled from a subring of the sample space. Then, we study the case where the secret is sampled from an ideal of the sample space or a coset thereof (equivalently, some of its CRT coordinates are fixed or leaked). In the latter, we show an interesting threshold phenomenon where the amount of RLWE noise determines whether the problem is tractable.Lastly, we address the long standing question of whether high-entropy secret is sufficient for RLWE to be intractable. Our result on sampling from ideals shows that simply requiring high entropy is insufficient. We therefore propose a broad class of distributions where we conjecture that hardness should hold, and provide evidence via reduction to a concrete lattice problem.

Fully homomorphic encryption schemes (FHE) allow to apply arbitrary efficient computation to encrypted data without decrypting it first. In Quantum FHE (QFHE) we may want to apply an arbitrary quantumly efficient computation to (classical or quantum) encrypted data.We present a QFHE scheme with classical key generation (and classical encryption and decryption if the encrypted message is itself classical) with comparable properties to classical FHE. Security relies on the hardness of the learning with errors (LWE) problem with polynomial modulus, which translates to the worst case hardness of approximating short vector problems in lattices to within a polynomial factor. Up to polynomial factors, this matches the best known assumption for classical FHE. Similarly to the classical setting, relying on LWE alone only implies leveled QFHE (where the public key length depends linearly on the maximal allowed evaluation depth). An additional circular security assumption is required to support completely unbounded depth. Interestingly, our circular security assumption is the same assumption that is made to achieve unbounded depth multi-key classical FHE.Technically, we rely on the outline of Mahadev (arXiv 2017) which achieves this functionality by relying on super-polynomial LWE modulus and on a new circular security assumption. We observe a connection between the functionality of evaluating quantum gates and the circuit privacy property of classical homomorphic encryption. While this connection is not sufficient to imply QFHE by itself, it leads us to a path that ultimately allows using classical FHE schemes with polynomial modulus towards constructing QFHE with the same modulus.

The hardness of the learning with errors (LWE) problem is one of the most fruitful resources of modern cryptography. In particular, it is one of the most prominent candidates for secure post-quantum cryptography. Understanding its quantum complexity is therefore an important goal.We show that under quantum polynomial time reductions, LWE is equivalent to a relaxed version of the dihedral coset problem (DCP), which we call extrapolated DCP (eDCP). The extent of extrapolation varies with the LWE noise rate. By considering different extents of extrapolation, our result generalizes Regev’s famous proof that if DCP is in BQP (quantum poly-time) then so is LWE (FOCS 02). We also discuss a connection between eDCP and Childs and Van Dam’s algorithm for generalized hidden shift problems (SODA 07).Our result implies that a BQP solution for LWE might not require the full power of solving DCP, but rather only a solution for its relaxed version, eDCP, which could be easier.

We show that any multi-party functionality can be evaluated using a two-round protocol with perfect correctness and perfect semi-honest security, provided that the majority of parties are honest. This settles the round complexity of information-theoretic semi-honest MPC, resolving a longstanding open question (cf. Ishai and Kushilevitz, FOCS 2000). The protocol is efficient for $${\mathrm {NC}}^1$$NC1 functionalities. Furthermore, given black-box access to a one-way function, the protocol can be made efficient for any polynomial functionality, at the cost of only guaranteeing computational security.Technically, we extend and relax the notion of randomized encoding to specifically address multi-party functionalities. The property of a multi-party randomized encoding (MPRE) is that if the functionality g is an encoding of the functionality f, then for any (permitted) coalition of players, their respective outputs and inputs in g allow them to simulate their respective inputs and outputs in f, without learning anything else, including the other outputs of f.

We construct a two-message oblivious transfer (OT) protocol without setup that guarantees statistical privacy for the sender even against malicious receivers. Receiver privacy is game based and relies on the hardness of learning with errors (LWE). This flavor of OT has been a central building block for minimizing the round complexity of witness indistinguishable and zero knowledge proof systems, non-malleable commitment schemes and multi-party computation protocols, as well as for achieving circuit privacy for homomorphic encryption in the malicious setting. Prior to this work, all candidates in the literature from standard assumptions relied on number theoretic assumptions and were thus insecure in the post-quantum setting. This work provides the first (presumed) post-quantum secure candidate and thus allows to instantiate the aforementioned applications in a post-quantum secure manner.Technically, we rely on the transference principle: Either a lattice or its dual must have short vectors. Short vectors, in turn, can be translated to information loss in encryption. Thus encrypting one message with respect to the lattice and one with respect to its dual guarantees that at least one of them will be statistically hidden.