Smart contracts are self-executing agreements written in program code and are envisioned to be one of the main applications of blockchain technology. While they are supported by prominent cryptocurrencies such as Ethereum, their further adoption is hindered by fundamental scalability challenges. For instance, in Ethereum contract execution suffers from a latency of more than 15 seconds, and the total number of contracts that can be executed per second is very limited. State channel networks are one of the core primitives aiming to address these challenges. They form a second layer over the slow and expensive blockchain, thereby enabling instantaneous contract processing at negligible costs. In this work we present the first complete description of a state channel network that exhibits the following key features. First, it supports virtual multi-party state channels, i.e. state channels that can be created and closed without blockchain interaction and that allow contracts with any number of parties. Second, the worst case time complexity of our protocol is constant for arbitrary complex channels. This is in contrast to the existing virtual state channel construction that has worst case time complexity linear in the number of involved parties. In addition to our new construction, we provide a comprehensive model for the modular design.

A witness encryption (WE) scheme can take any NP statement as a public-key and use it to encrypt a message. If the statement is true then it is possible to decrypt the message given a corresponding witness, but if the statement is false then the message is computationally hidden. Ideally, the encryption procedure should run in polynomial time, but it is also meaningful to define a weaker notion, which we call non-trivially exponentially efficient WE (XWE), where the encryption run-time is only required to be much smaller than the trivial $2^{m}$ bound for NP relations with witness size $m$. We show how to construct such XWE schemes for all of NP with encryption run-time $2^{m/2}$ under the sub-exponential learning with errors (LWE) assumption. For NP relations that can be verified in NC1 (e.g., SAT) we can also construct such XWE schemes under the sub-exponential Decisional Bilinear Diffie-Hellman (DBDH) assumption. Although we find the result surprising, it follows via a very simple connection to attribute-based encryption. We also show how to upgrade the above results to get non-trivially exponentially efficient indistinguishability obfuscation for null circuits (niO), which guarantees that the obfuscations of any two circuits that always output 0 are indistinguishable. In particular, under the LWE assumptions we get a XniO scheme where the obfuscation time is $2^{n/2}$ for all circuits with input size $n$. It is known that in the case of indistinguishability obfuscation (iO) for all circuits, non-trivially efficient XiO schemes imply fully efficient iO schemes (Lin et al., PKC '16) but it remains as a fascinating open problem whether any such connection exists for WE or niO. Lastly, we explore a potential approach toward constructing fully efficient WE and niO schemes via multi-input ABE.

Though it is well known that the roots of any affine polynomial over finite field can be computed by a system of linear equations by using a normal base of the field, such solving approach appears to be difficult to apply when the field is fairly large. Thus, it may be of great interest to find explicit representation of the solutions independently of the field base. This was previously done only for quadratic equations over binary finite field. This paper gives explicit representation of solutions for much wider class of affine polynomials over binary prime field.

Non-malleable codes are encoding schemes that provide protections against various classes of tampering attacks. Recently Faust et al. (CRYPTO 2017) initiated the study of space- bounded non-malleable codes that provide such protections against tampering within small- space devices. They put forward a construction based on any non-interactive proof-of-space (NIPoS). However, the scheme only protects against an a priori bounded number of tampering attacks. We construct non-malleable codes that are resilient to an unbounded polynomial number of space-bounded tamperings. Towards that we introduce a stronger variant of NIPoS called proof-extractable NIPoS (PExt-NIPoS), and propose two approaches of constructing such a primitive. Using a new proof strategy we show that the generic encoding scheme of Faust et al. achieves unbounded tamper-resilience when instantiated with a PExt-NIPoS. We show two methods to construct PExt-NIPoS: 1. The first method uses a special family of “memory-hard” graphs, called challenge-hard graphs (CHG), a notion we introduce here. We instantiate such family of graphs based on an extension of stack of localized expanders (first used by Ren and Devadas in the context of proof-of-space). In addition, we show that the graph construction used as a building block for the proof-of-space by Dziembowski et al. (CRYPTO 2015) satisfies challenge-hardness as well. These two CHG-instantiations lead to continuous space-bounded NMC with different features in the random oracle model. 2. Our second instantiation relies on a new measurable property, called uniqueness of NIPoS. We show that standard extractability can be upgraded to proof-extractability if the NIPoS also has uniqueness. We propose a simple heuristic construction of NIPoS, that achieves (partial) uniqueness, based on a candidate memory-hard function in the standard model and a publicly verifiable computation with small-space verification. Instantiating the encoding scheme of Faust et al. with this NIPoS, we obtain a continuous space-bounded NMC that supports the “most practical” parameters, complementing the provably secure but “relatively impractical” CHG-based constructions. Additionally, we revisit the construction of Faust et al. and observe that due to the lack of uniqueness of their NIPoS, the resulting encoding schemes yield “highly impractical” parameters in the continuous setting. We conclude the paper with a comparative study of all our non-malleable code constructions with an estimation of concrete parameters.

The Fiat-Shamir heuristic transforms a public-coin interactive proof into a non-interactive argument, by replacing the verifier with a cryptographic hash function that is applied to the protocol’s transcript. Constructing hash functions for which this transformation is sound is a central and long-standing open question in cryptography. We show that solving the End-of-Metered-Line problem is no easier than breaking the soundness of the Fiat-Shamir transformation when applied to the sumcheck protocol. In particular, if the transformed protocol is sound, then any hard problem in #P gives rise to a hard distribution in the class CLS, which is contained in PPAD. Our main technical contribution is a stateful incrementally verifiable procedure that, given a SAT instance over n variables, counts the number of satisfying assignments. This is accomplished via an exponential sequence of small steps, each computable in time poly(n). Incremental verifiability means that each intermediate state includes a sumcheck-based proof of its correctness, and the proof can be updated and verified in time poly(n). Combining our construction with a hash family proposed by Canetti et al. [STOC 2019] gives rise to a distribution in the class CLS, which is provably hard under the assumption that any one of a class of fully homomorphic encryption (FHE) schemes has almost-optimal security against quasi-polynomial time adversaries, and under the additional worst-case assumption that there is no polynomial time algorithm for counting the number of satisfying assignments for formulas over a polylogarithmic number of variables.

It has been 70 years since the publication of the seminal outstanding work of Claude Elwood Shannon, in which he first gave a mathematical definition of the cryptosystem and introduced the concept of perfect ciphers. He also examined the conditions in which such a ciphers exist. Shannon's results in one form or another are presented in almost all books on cryptography. One of his result deals with so-called endomorphic ciphers in which the cardinalities of the message space $\mathcal{M}$ and the ciphertexts $\mathcal{C}$ are the same. The Vernam cipher (one-time pad) is the most famous representative of such ciphers. Moreover, it's the only one known to be perfect. Surprisingly, we have found a mistake in the Shannon's result. Namely, Shannon stated that an endomorphic cipher, in which the keyspace $\mathcal{K}$ has the same cardinality as message space, is perfect if and only if two conditions are satisfied. The first suggests that for any pair plaintext - ciphertext there exists only one key that translates this plaintext into this ciphertext. The second argues that the key distribution must be uniform. We show, that these two conditions are not really enough. We prove in three different ways that the plaintexts must also be equally probable. Moreover, we study the general endomorphic cipher and get the same result. It follows, that in practice perfect endomorphic ciphers do not exist.

The era of PUFs has been characterized by the efforts put into research and the development of PUFs that are robust against attacks, in particular, machine learning (ML) attacks. In the lack of systematic and provable methods for this purpose, we have witnessed the ever-continuing competition between PUF designers/ manufacturers, cryptanalysts, and of course, adversaries that maliciously break the security of PUFs. This is despite a series of acknowledged principles developed in cryptography and complexity theory, under the umbrella term ``hardness amplification." The goal of studies on the hardness amplification is to build a strongly secure construction out of considerably weaker primitives. This paper aims at narrowing the gap between these studies and hardware security, specifically for applications in the domain of PUFs. To this end, we first review an example of practical efforts made to construct more secure PUFs, namely the concept of rolling PUFs. Based on what can be learned from this and central insights provided by the ML and complexity theory, we propose a new PUF-based scheme built around the idea of using a new function, namely, the Tribes function, which combines the outputs of a set of PUFs to generate the final response. Our theoretical findings are discussed in an exhaustive manner and supported by the results of experiments, conducted extensively on real-world PUFs.

We consider the normalized distribution of the overall running times of some cryptographic algorithms, and what information they reveal about the algorithms. Recent work of Deift, Menon, Olver, Pfrang, and Trogdon has shown that certain numerical algorithms applied to large random matrices exhibit a characteristic distribution of running times, which depends only on the algorithm but are independent of the choice of probability distributions for the matrices. Different algorithms often exhibit different running time distributions, and so the histograms for these running time distributions provide a "time-signature" for the algorithms, making it possible, in many cases, to distinguish one algorithm from another. In this paper we extend this analysis to cryptographic algorithms, and present examples of such algorithms with time-signatures that are indistinguishable, and others with time-signatures that are clearly distinct.

Murphy, Murky, Mopey, Moody, and Morose decide to write a paper together over the Internet and submit it to the prestigious CRYPTO’19 conference that has the most amazing PC. They encounter a few problems. First, not everyone is online every day: some are lazy and go skiing on Mondays; others cannot use git correctly and they are completely unaware that they are losing messages. Second, a small subset of the co-authors may be secretly plotting to disrupt the project (e.g., because they are writing a competing paper in stealth). Suppose that each day, sufficiently many honest co-authors are online (and use git correctly); moreover, suppose that messages checked into git on Monday can be correctly received by honest and online co-authors on Tuesday or any future day. Can the honest co-authors successfully finish the paper in a small number of days such that they make the CRYPTO deadline; and perhaps importantly, can all the honest co-authors, including even those who are lazy and those who sometimes use git incorrectly, agree on the final theorem?

We propose a generic framework for perfectly hiding UC-Commitment schemes in the Global Random Oracle model of Canetti \textit{el at.} (CCS 14). The main building block of our construction is a novel primitive called Sampleable-Range Trapdoor Function, that is, a trapdoor function for which there is a non-negligible probability of finding preimages when given a uniformly chosen element of its codomain and the corresponding trapdoor. To show the versatility of the framework, we give concrete instantiations based on factoring, code-based, and lattice-based hardness assumptions. Our construction yields the first lattice-based UC-Commitment scheme (not constructed via generic transformations, such as via Oblivious Transfer), and achieves what we call \textit{phase-adaptive security}, a novel security notion we introduce which is stronger than static security. Achieving adaptive security for UC-Commitment schemes is non-trivial and, usually, comes at the price of efficiency. Phase-adaptive security stands between adaptive and static security, and may be of independent interest. In this model, adversaries can corrupt at the beginning or between the commitment and opening phases of the protocol, but not during their execution. This new model is motivated by the fact that, in practice, it is more likely that parties are corrupted between phases of the protocol (where a relatively long period may elapse) than during their execution.

A sub-Gaussian distribution is any probability distribution that has tails bounded by a Gaussian and has a mean of zero. It is well known that the sum of independent sub-Gaussians is again sub-Gaussian. This note generalizes this result to sums of sub- Gaussians that may not be independent, under the assumption a certain conditional distribution is also sub-Gaussian. This general result is useful in the study of noise growth in (fully) homomorphic encryption schemes [CGHX19, CGGI17], and hopefully useful for other applications.

In this paper, we study the generic hardness of the inversion problem on a ring, which is a problem to compute the inverse of a given prime $c$ by just using additions, subtractions and multiplications on the ring. If the characteristic of an underlying ring is public and coprime to $c$, then it is easy to compute the inverse of $c$ by using the extended Euclidean algorithm. On the other hand, if the characteristic is hidden, it seems difficult to compute it. For discussing the generic hardness of the inversion problem, we first extend existing generic ring models to capture a ring of an unknown characteristic. Then we prove that there is no generic algorithm to solve the inversion problem in our model when the underlying ring is isomorphic to $\mathbb{Z}_p$ for a randomly chosen prime $p$ assuming the hardness of factorization of an unbalanced modulus. We also study a relation between the inversion problem on a ring and a self-bilinear map. We give a ring-based construction of a self-bilinear map, and prove that natural complexity assumptions including the multilinear computational Diffie-Hellman (MCDH) assumption hold w.r.t the resulting sef-bilinear map if the inversion problem is hard on the underlying ring.

We design, implement, and evaluate a zkSNARK for Rank-1 Constraint Satisfaction (R1CS), a widely-deployed NP-complete language that is undergoing standardization. Our construction uses a transparent setup, is plausibly post-quantum secure, and uses lightweight cryptography. A proof attesting to the satisfiability of n constraints has size $O(\log^2 n)$; it can be produced with $O(n \log n)$ field operations and verified with $O(n)$. At 128 bits of security, proofs are less than 130kB even for several million constraints, more than 20x shorter than prior zkSNARK with similar features. A key ingredient of our construction is a new Interactive Oracle Proof (IOP) for solving a *univariate* analogue of the classical sumcheck problem [LFKN92], originally studied for *multivariate* polynomials. Our protocol verifies the sum of entries of a Reed--Solomon codeword over any subgroup of a field. We also provide libiop, an open-source library for writing IOP-based arguments, in which a toolchain of transformations enables programmers to write new arguments by writing simple IOP sub-components. We have used this library to specify our construction and prior ones.

We study protocols that rely on a public ledger infrastructure, concentrating on protocols for zero-knowledge contingent payment, whose security properties combine diverse notions of fairness and privacy. We argue that rigorous models are required for capturing the ledger semantics, the protocol-ledger interaction, the cryptographic primitives and, ultimately, the security properties one would like to achieve. Our focus is on a particular level of abstraction, where network messages are represented by a term algebra, protocol execution by state transition systems (e.g. multiset rewrite rules) and where the properties of interest can be analyzed with automated verification tools. We propose models for: (1) the rules guiding the ledger execution, taking the coin functionality of public ledgers such as Bitcoin as an example; (2) the security properties expected from ledger-based zero-knowledge contingent payment protocols; (3) two different security protocols that aim at achieving these properties relying on different ledger infrastructures; (4) reductions that allow simpler term algebras for homomorphic cryptographic schemes. Altogether, these models allow us to derive a first automated verification for ledger-based zero-knowledge contingent payment using the Tamarin prover. Furthermore, our models help in clarifying certain underlying assumptions, security and efficiency tradeoffs that should be taken into account when deploying protocols on the blockchain.

We introduce a new cryptographic primitive called laconic function evaluation (LFE). Using LFE, Alice can compress a large circuit $f$ into a small digest. Bob can encrypt some data $x$ under this digest in a way that enables Alice to recover $f(x)$ without learning anything else about Bob's data. For the scheme to be laconic, we require that the size of the digest, the run-time of the encryption algorithm and the size of the ciphertext should all be small, much smaller than the circuit-size of $f$. We construct an LFE scheme for general circuits under the learning with errors (LWE) assumption, where the above parameters only grow polynomially with the depth but not the size of the circuit. We then use LFE to construct secure 2-party and multi-party computation (2PC, MPC) protocols with novel properties: * We construct a 2-round 2PC protocol between Alice and Bob with respective inputs $x_A,x_B$ in which Alice learns the output $f(x_A,x_B)$ in the second round. This is the first such protocol which is "Bob-optimized", meaning that Alice does all the work while Bob's computation and the total communication of the protocol are smaller than the size of the circuit $f$ or even Alice's input $x_A$. In contrast, prior solutions based on fully homomorphic encryption are "Alice-optimized". * We construct an MPC protocol, which allows $N$ parties to securely evaluate a function $f(x_1,...,x_N)$ over their respective inputs, where the total amount of computation performed by the parties during the protocol execution is smaller than that of evaluating the function itself! Each party has to individually pre-process the circuit $f$ before the protocol starts and post-process the protocol transcript to recover the output after the protocol ends, and the cost of these steps is larger than the circuit size. However, this gives the first MPC where the computation performed by each party during the actual protocol execution, from the time the first protocol message is sent until the last protocol message is received, is smaller than the circuit size.

A fundamental problem in the theory of secure multi-party computation (MPC) is to characterize functions with more than 2 parties which admit MPC protocols with information-theoretic security against passive corruption. This question has seen little progress since the work of Chor and Ishai (1996), which demonstrated difficulties in resolving it. In this work, we make significant progress towards resolving this question in the important case of aggregating functionalities, in which m parties P1, . . . , Pm hold inputs x1, . . . , xm and an aggregating party P0 must learn f(x1,...,xm). We uncover a rich class of algebraic structures that are closely related to secure computability, namely, “Commuting Permutations Systems” (CPS) and its variants. We present an extensive set of results relating these algebraic structures among themselves and to MPC, including new protocols, impossibility results and separations. Our results include a necessary algebraic condition and slightly stronger sufficient algebraic condition for a function to admit information-theoretically secure MPC protocols. We also introduce and study new models of minimally interactive MPC (called UNIMPC and UNIMPC*), which not only help in understanding our positive and negative results better, but also open up new avenues for studying the cryptographic complexity landscape of multi-party functionalities. Our positive results include novel protocols in these models, which may be of independent practical interest. Finally, we extend our results to a definition that requires UC security as well as semi-honest security (which we term strong security). In this model we are able to carry out the characterization of all computable functions, except for a gap in the case of aggregating functionalities.

We construct a verifable delay function (VDF) by showing how the Rivest-Shamir-Wagner time-lock puzzle can be made publicly verifiable. Concretely, we give a statistically sound public-coin protocol to prove that a tuple $(N,x,T,y)$ satisfies $y=x^{2^T}\pmod N$ where the prover doesn't know the factorization of $N$ and its running time is dominated by solving the puzzle, that is, compute $x^{2^T}$, which is conjectured to require $T$ sequential squarings. To get a VDF we make this protocol non-interactive using the Fiat-Shamir heuristic. The motivation for this work comes from the Chia blockchain design, which uses a VDF as a key ingredient. For typical parameters ($T\le 2^{40},N=2048$), our proofs are of size around $10KB$, verification cost around three RSA exponentiations and computing the proof is $8000$ times faster than solving the puzzle even without any parallelism.

Non-malleable codes, introduced by Dziembowski, Pietrzak and Wichs in ICS 2010, have emerged in the last few years as a fundamental object at the intersection of cryptography and coding theory. Non-malleable codes provide a useful message integrity guarantee in situations where traditional error-correction (and even error-detection) is impossible; for example, when the attacker can completely overwrite the encoded message. Informally, a code is non-malleable if the message contained in a modified codeword is either the original message, or a completely ``unrelated value''. Although such codes do not exist if the family of ``tampering functions'' {\mathcal F} allowed to modify the original codeword is completely unrestricted, they are known to exist for many broad tampering families {\mathcal F}. The family which received the most attention is the family of tampering functions in the so called (2-part) {\em split-state} model: here the message x is encoded into two shares L and R, and the attacker is allowed to arbitrarily tamper with each L and R individually. Dodis, Kazana, and the authors in STOC 2015 developed a generalization of non-malleable codes called the concept of non-malleable reduction, where a non-malleable code for a tampering family {\mathcal F} can be seen as a non-malleable reduction from {\mathcal F} to a family NM of functions comprising the identity function and constant functions. They also gave a constant-rate reduction from a split-state tampering family to a tampering family {\mathcal G} containing so called $2$-lookahead functions, and forgetful functions. In this work, we give a constant rate non-malleable reduction from the family {\mathcal G} to NM, thereby giving the first {\em constant rate non-malleable code in the split-state model.} Central to our work is a technique called inception coding which was introduced by Aggarwal, Kazana and Obremski in TCC 2017, where a string that detects tampering on a part of the codeword is concatenated to the message that is being encoded.

While digital secret keys appear indispensable in modern cryptography and security, they also routinely constitute a main attack point of the resulting hardware systems. Some recent approaches have tried to overcome this problem by simply avoiding keys and secrets in vulnerable systems. To start with, physical unclonable functions (PUFs) have demonstrated how “classical keys”, i.e., permanently stored digital secret keys, can be evaded, realizing security devices that might be called “classically key-free”. Still, most PUFs induce certain types of physical secrets deep in the hardware, whose disclosure to adversaries breaks security as well. Examples include the manufacturing variations that determine the power-up states of SRAM PUFs, or the signal runtimes of Arbiter PUFs, both of which have been extracted from PUF-hardware in practice, breaking security. A second generation of physical security primitives, such a SIMPLs/PPUFs and Unique Objects, recently has shown promise to overcome this issue, however. Perhaps counterintuitively, they would enable completely “secret-free” hardware, where adversaries might inspect every bit and atom, and learn any information present in any form in the hardware, without being able to break security. This concept paper takes this situation as starting point, and categorizes, formalizes, and surveys the currently emerging areas of key-free and, more importantly, secret-free security. Our treatment puts keys, secrets, and their respective avoidance into the center of the currently emerging physical security methods. It so aims to lay the foundations for future, secret-free security hardware, which would be innately and provably immune against any physical probing and key extraction.

We construct efficient, unconditional non-malleable codes that are secure against tampering functions computed by decision trees of depth $d = n^{1/4-o(1)}$. In particular, each bit of the tampered codeword is set arbitrarily after adaptively reading up to $d$ arbitrary locations within the original codeword. Prior to this work, no efficient unconditional non-malleable codes were known for decision trees beyond depth $O(\log^2 n)$. Our result also yields efficient, unconditional non-malleable codes that are $\exp(-n^{\Omega(1)})$-secure against constant-depth circuits of $\exp(n^{\Omega(1)})$-size. Prior work of Chattopadhyay and Li (STOC 2017) and Ball et al. (FOCS 2018) only provide protection against $\exp(O(\log^2n))$-size circuits with $\exp(-O(\log^2n))$-security. We achieve our result through simple non-malleable reductions of decision tree tampering to split-state tampering. As an intermediary, we give a simple and generic reduction of leakage-resilient split-state tampering to split-state tampering with improved parameters. Prior work of Aggarwal et al. (TCC 2015) only provides a reduction to split-state non-malleable codes with decoders that exhibit particular properties.

In this work, we study the intrinsic complexity of black-box Universally Composable (UC) secure computation based on general assumptions. We present a thorough study in various corruption modelings while focusing on achieving security in the common reference string (CRS) model. Our results involve the following: 1. Static UC secure computation. Designing the first static UC oblivious transfer protocol based on public-key encryption and stand-alone semi-honest oblivious transfer. As a corollary we obtain the first black-box constructions of UC secure computation assuming only two-round semi-honest oblivious transfer. 2. One-sided UC secure computation. Designing adaptive UC two-party computation with single corruptions assuming public-key encryption with oblivious ciphertext generation. 3. Adaptive UC secure computation. Designing adaptively secure UC commitment scheme assuming only public-key encryption with oblivious ciphertext generation. As a corollary we obtain the first black-box constructions of adaptive UC secure computation assuming only (trapdoor) simulatable public-key encryption (as well as a variety of concrete assumptions). We remark that such a result was not known even under non-black-box constructions.

One of the most important tools for assessing hardness assumptions in cryptography is the Generic Group Model (GGM). Over the past two decades, numerous assumptions have been analyzed within this model. While a proof in the GGM can certainly provide some measure of confidence in an assumption, its scope is rather limited since it does not capture group-specific algorithms that make use of the representation of the group. To overcome this limitation, we propose the Algebraic Group Model (AGM), a model that lies in between the standard model and the GGM. It is the first restricted model of computation covering group-specific algorithms yet allowing to derive simple and meaningful security statements. We show that several important assumptions, among them the Computational Diffie-Hellman, the Strong Diffie-Hellman, and the interactive LRSW assumptions, are equivalent to the Discrete Logarithm (DLog) assumption in the AGM. On the more practical side, we prove tight security reductions for two important schemes in the AGM to DLog or a variant thereof: the BLS signature scheme and Groth's zero-knowledge SNARK (Eurocrypt '16), which is the most efficient SNARK for which only a proof in the GGM was known. Moreover, in combination with known lower bounds on the Discrete Logarithm assumption in the GGM, our results can be used to derive lower bounds for all the above-mentioned results in the GGM.

We present a construction of an adaptively single-key secure constrained PRF (CPRF) for $\mathbf{NC}^1$ assuming the existence of indistinguishability obfuscation (IO) and the subgroup hiding assumption over a (pairing-free) composite order group. This is the first construction of such a CPRF in the standard model without relying on a complexity leveraging argument. To achieve this, we first introduce the notion of partitionable CPRF, which is a CPRF accommodated with partitioning techniques and combine it with shadow copy techniques often used in the dual system encryption methodology. We present a construction of partitionable CPRF for $\mathrm{NC}^1$ based on IO and the subgroup hiding assumption over a (pairing-free) group. We finally prove that an adaptively single-key secure CPRF for $\mathbf{NC}^1$ can be obtained from a partitionable CPRF for $\mathbf{NC}^1$ and IO.

Motivated by the quest for scalable and succinct zero knowledge arguments, we revisit worst-case-to-average-case reductions for linear spaces, raised by [Rothblum, Vadhan, Wigderson, STOC 2013]. The previous state of the art by [Ben-Sasson, Kopparty, Saraf, CCC 2018] showed that if some member of an affine space $U$ is $\delta$-far in relative Hamming distance from a linear code $V$ — this is the worst-case assumption — then most elements of $U$ are almost-$\delta$-far from $V$ — this is the average case. However, this result was known to hold only below the “double Johnson” function of the relative distance $\delta_V$ of the code $V$ , i.e., only when $\delta < 1 - \sqrt[4]{1 - \delta_V}$. First, we increase the soundness-bound to the “one-and-a-half Johnson” function of $\delta_V$ and show that the average distance of $U$ from $V$ is nearly $\delta$ for any worst-case distance $\delta$ smaller than $1 - \sqrt[3]{1 - \delta_V}$. This bound is tight, which is somewhat surprising because the one-and-a-half Johnson function is unfamiliar in the literature on error correcting codes. To improve soundness further for Reed Solomon codes we sample outside the box. We suggest a new protocol in which the verifier samples a single point $z$ outside the box $D$ on which codewords are evaluated, and asks the prover for the value at $z$ of the interpolating polynomial of a random element of $U$. Intuitively, the answer provided by the prover “forces” it to choose one codeword from a list of “pretenders” that are close to $U$. We call this technique Domain Extending for Eliminating Pretenders (DEEP). The DEEP method improves the soundness of the worst-case-to-average-case reduction for RS codes up their list decoding radius. This radius is bounded from below by the Johnson bound, implying average distance is approximately $\delta$ for all $\delta < 1 - \sqrt{1 - \delta_V}$. Under a plausible conjecture about the list decoding radius of Reed-Solomon codes, average distance from $V$ is approximately $\delta$ for all $\delta$. The DEEP technique can be generalized to all linear codes, giving improved reductions for capacity-achieving list-decodable codes. Finally, we use the DEEP technique to devise two new protocols: • An Interactive Oracle Proof of Proximity (IOPP) for RS codes, called DEEP-FRI. This soundness of the protocol improves upon that of the FRI protocol of [Ben-Sasson et al., ICALP 2018] while retaining linear arithmetic proving complexity and logarithmic verifier arithmetic complexity. • An Interactive Oracle Proof (IOP) for the Algebraic Linking IOP (ALI) protocol used to construct zero knowledge scalable transparent arguments of knowledge (ZK-STARKs) in [Ben-Sasson et al., eprint 2018]. The new protocol, called DEEP-ALI, improves soundness of this crucial step from a small constant $< 1/8$ to a constant arbitrarily close to $1$.

Abstract. This work describes a fast fully homomorphic encryption scheme over the torus (TFHE), that revisits, generalizes and improves the fully homomorphic encryption (FHE) based on GSW and its ring variants. The simplest FHE schemes consist in bootstrapped binary gates. In this gate bootstrapping mode, we show that the scheme FHEW of [29] can be expressed only in terms of external product between a GSW and a LWE ciphertext. As a consequence of this result and of other optimizations, we decrease the running time of their bootstrapping from 690ms to 13ms single core, using 16MB bootstrapping key instead of 1GB, and preserving the security parameter. In leveled homomorphic mode, we propose two methods to manipulate packed data, in order to decrease the ciphertext expansion and to optimize the evaluation of look-up tables and arbitrary functions in RingGSW based homomorphic schemes. We also extend the automata logic, introduced in [31], to the efficient leveled evaluation of weighted automata, and present a new homomorphic counter called TBSR, that supports all the elementary operations that occur in a multiplication. These improvements speed-up the evaluation of most arithmetic functions in a packed leveled mode, with a noise overhead that remains additive. We finally present a new circuit bootstrapping that converts LWE ciphertexts into low-noise RingGSW ciphertexts in just 137ms, which makes the leveled mode of TFHE composable, and which is fast enough to speed-up arithmetic functions, compared to the gate bootstrapping approach. Finally, we provide an alternative practical analysis of LWE based schemes, which directly relates the security parameter to the error rate of LWE and the entropy of the LWE secret key, and we propose concrete parameter sets and timing comparison for all our constructions.

We show that every construction of one-time signature schemes from a random oracle achieves black-box security at most $2^{(1+o(1))q}$, where $q$ is the total number of oracle queries asked by the key generation, signing, and verification algorithms. That is, any such scheme can be broken with probability close to $1$ by a (computationally unbounded) adversary making $2^{(1+o(1))q}$ queries to the oracle. This is tight up to a constant factor in the number of queries, since a simple modification of Lamport's one-time signatures (Lamport'79) achieves $2^{(0.812-o(1))q}$ black-box security using $q$ queries to the oracle. Our result extends (with a loss of a constant factor in the number of queries) also to the random permutation and ideal-cipher oracles. Since the symmetric primitives (e.g. block ciphers, hash functions, and message authentication codes) can be constructed by a constant number of queries to the mentioned oracles, as corollary we get lower bounds on the efficiency of signature schemes from symmetric primitives when the construction is black-box. This can be taken as evidence of an inherent efficiency gap between signature schemes and symmetric primitives.

For all vectorial boolean functions up to dimension 4, we present canonical representatives for all extended affine (EA) and CCZ equivalence classes. We include the size of each class, as well as its algebraic degree and extended Walsh spectrum. We also answer the following questions: How large are these classes? Which of these classes contain bijective functions? And how are these classes grouped into CCZ equivalence classes?

The need for high-quality randomness in cryptography makes random-number generation one of its most fundamental tasks. A recent important line of work (initiated by Dodis et al., CCS ’13) focuses on the notion of *robustness* for *pseudorandom number generators (PRNGs) with inputs*—these are primitives that use various sources to accumulate sufficient entropy into a state, from which pseudorandom bits are extracted. Robustness ensures that PRNGs remain secure even under state compromise and adversarial control of entropy sources. However, the achievability of robustness inherently depends on a seed, or, alternatively, on an ideal primitive (e.g., a random oracle), independent of the source of entropy. Both assumptions are problematic: seed generation requires randomness to start with, and it is arguable whether the seed or the ideal primitive can be kept independent of the source. This paper resolves this dilemma by putting forward new notions of robustness which enable both (1) *seedless* PRNGs and (2) *primitive-dependent* adversarial sources of entropy. To bypass obvious impossibility results, we make a realistic compromise by requiring that the source produce sufficient entropy even given its evaluations of the underlying primitive. We also provide natural, practical, and provably secure constructions based on hash-function designs from compression functions, block ciphers, and permutations. Our constructions can be instantiated with minimal changes to industry-standard hash functions SHA-2 and SHA-3, or HMAC (as used for the key derivation function HKDF), and can be downgraded to *(online) seedless randomness extractors*, which are of independent interest. On the way we consider both a *computational* variant of robustness, where attackers only make a bounded number of queries to the ideal primitive, as well as a new *information-theoretic* variant, which dispenses with this assumption to a certain extent, at the price of requiring a high rate of injected weak randomness (as it is, e.g., plausible on Intel’s on-chip RNG). The latter notion enables applications such as everlasting security. Finally, we show that the CBC extractor, used by Intel’s on-chip RNG, is provably insecure in our model.

Let $N_a$ be the number of solutions to the equation $x^{2^k+1}+x+a=0$ in $\mathbb F_{n}$ where $\gcd(k,n)=1$. In 2004, by Bluher it was known that possible values of $N_a$ are only 0, 1 and 3. In 2008, Helleseth and Kholosha have got criteria for $N_a=1$ and an explicit expression of the unique solution when $\gcd(k,n)=1$. In 2014, Bracken, Tan and Tan presented a criterion for $N_a=0$ when $n$ is even and $\gcd(k,n)=1$. This paper completely solves this equation $x^{2^k+1}+x+a=0$ with only condition $\gcd(n,k)=1$. We explicitly calculate all possible zeros in $\mathbb F_{n}$ of $P_a(x)$. New criterion for which $a$, $N_a$ is equal to $0$, $1$ or $3$ is a by-product of our result.

In the past 15 years, cryptography has made considerable progress in expanding the adversarial attack model to cover side-channel attacks, and has built schemes to provably defend against some of them. This survey covers the main models and results in this so-called "leakage-resilient" cryptography.

In this note, we show that a strong notion of KDM security cannot be obtained by any encryption scheme in the auxiliary input setting, assuming Learning With Errors (LWE) and one-way permutations. The notion of security we deal with guarantees that for any (possibly inefficient) function $f$, it is computationally hard to distinguish between an encryption of 0s and an encryption of f(pk, z), where pk is the public key and z is the auxiliary input. Furthermore, we show that this holds even when restricted to bounded-length auxiliary input where z is much shorter than pk under the additional assumption that (non-leveled) fully homomorphic encryption exists.

Any $d$-regular graph on $n$ vertices with spectral expansion $\lambda$ satisfying $n = \Omega(d^3\log(d)/\lambda)$ yields a $O\left(\frac{\lambda^{3/2}}{d}\right)$-non-malleable code for single-bit messages in the split-state model.

Boolean functions used in some cryptosystems of stream ciphers should satisfy various criteria simultaneously to resist some known attacks. The fast algebraic attack (FAA) is feasible if one can find a nonzero function $g$ of low algebraic degree and a function $h$ of algebraic degree significantly lower than $n$ such that $f\cdot g=h$. Then one new cryptographic property fast algebraic immunity was proposed, which measures the ability of Boolean functions to resist FAAs. It is a great challenge to determine the exact values of the fast algebraic immunity of an infinite class of Boolean functions with optimal algebraic immunity. In this letter, we explore the exact fast algebraic immunity of two subclasses of the majority function.

A geometry is a measure of restraint over the allowed 0.5n(n-1) distances between a set of n points (e.g. the metric and topological spaces). So defined, geometries lead to associated algebra. The complexities of such algebras are used to build cryptographic primitives. We propose then to push geometries to the limit -- unbound geometries -- where any two points may be assigned an arbitrary distance value, which may reflect a planning process or a randomized assignment. Regarding these distances as a cryptographic key, one could use the resultant algebras to carry out cryptographic missions. We define the mathematical framework for this aim, then present a few cryptographic primitives. Most effective implementation is through the new technology for “rock of randomness” establishing random distances through 3D printed molecular compounds. Security is proportional to the size of the ‘rock’. We use the term SpaceFlip to collectively refer to the unbound geometry, its associated algebra and the cryptographic tools derived from it.

We give a new proof of the existence of $O(n^{\epsilon})$-round public-coin concurrent zero-knowledge arguments for NP, where $\epsilon>0$ is an arbitrary constant. The security is proven in the plain model under the assumption that collision-resistant hash functions exist. (The existence of such concurrent zero-knowledge arguments was previously proven by Goyal (STOC'13) in the plain model under the same assumption.) In the proof, we use a new variant of the non-black-box simulation technique of Barak (FOCS'01). An important property of our simulation technique is that the simulator runs in a "straight-line" manner in the fully concurrent setting. Compared with the simulation technique of Goyal, which also has such a property, the analysis of our simulation technique is (arguably) simpler.

We observe that any key agreement protocol satisfying perfect completeness, regardless of its round complexity, can be used to construct a non-interactive commitment scheme. This observation simplifies the cryptographic assumptions required for some protocols that utilize non-interactive commitments and removes the need for ad-hoc constructions of non-interactive commitments from specific assumptions such as Learning with Errors.

Non-malleable Codes give us the following property: their codewords cannot be tampered into codewords of related messages. Privacy Amplification allows parties to convert their weak shared secret into a fully hidden, uniformly distributed secret key, while communicating on a fully tamperable public channel. In this work, we show how to construct a constant round privacy amplification protocol from any augmented split-state non-malleable code. Existentially, this gives us another primitive (in addition to optimal non-malleable extractors) whose optimal construction would solve the long-standing open problem of building constant round privacy amplification with optimal entropy loss. Instantiating our code with the current best known NMC gives us an $8$-round privacy amplification protocol with entropy loss $O(\log(n)+ \kappa \log (\kappa))$ and min-entropy requirement $\Omega(\log(n) +\kappa\log (\kappa))$, where $\kappa$ is the security parameter and $n$ is the length of the shared weak secret. In fact, for our result, even the weaker primitive of Non-malleable Randomness Encoders suffice. We view our result as an exciting connection between two of the most fascinating and well-studied information theoretic primitives, non-malleable codes and privacy amplification.

The Fiat-Shamir transformation is a useful approach to building non-interactive arguments (of knowledge) in the random oracle model. Unfortunately, existing proof techniques are incapable of proving the security of Fiat-Shamir in the quantum setting. The problem stems from (1) the difficulty of quantum rewinding, and (2) the inability of current techniques to adaptively program random oracles in the quantum setting. In this work, we show how to overcome the limitations above in many settings. In particular, we give mild conditions under which Fiat-Shamir is secure in the quantum setting. As an application, we show that existing lattice signatures based on Fiat-Shamir are secure without any modifications.

There is surprisingly little consensus on the precise role of the generator g in group-based assumptions such as DDH. Some works consider g to be a fixed part of the group description, while others take it to be random. We study this subtle distinction from a number of angles. - In the generic group model, we demonstrate the plausibility of groups in which random-generator DDH (resp. CDH) is hard but fixed-generator DDH (resp. CDH) is easy. We observe that such groups have interesting cryptographic applications. - We find that seemingly tight generic lower bounds for the Discrete-Log and CDH problems with preprocessing (Corrigan-Gibbs and Kogan, Eurocrypt 2018) are not tight in the sub-constant success probability regime if the generator is random. We resolve this by proving tight lower bounds for the random generator variants; our results formalize the intuition that using a random generator will reduce the effectiveness of preprocessing attacks. - We observe that DDH-like assumptions in which exponents are drawn from low-entropy distributions are particularly sensitive to the fixed- vs. random-generator distinction. Most notably, we discover that the Strong Power DDH assumption of Komargodski and Yogev (Komargodski and Yogev, Eurocrypt 2018) used for non-malleable point obfuscation is in fact false precisely because it requires a fixed generator. In response, we formulate an alternative fixed-generator assumption that suffices for a new construction of non-malleable point obfuscation, and we prove the assumption holds in the generic group model. We also give a generic group proof for the security of fixed-generator, low-entropy DDH (Canetti, Crypto 1997).

The quantum random oracle model (QROM) has become the standard model in which to prove the post-quantum security of random-oracle-based constructions. Unfortunately, none of the known proof techniques allow the reduction to record information about the adversary's queries, a crucial feature of many classical ROM proofs, including all proofs of indifferentiability for hash function domain extension. In this work, we give a new QROM proof technique that overcomes this ``recording barrier''. Our central observation is that when viewing the adversary's query and the oracle itself in the Fourier domain, an oracle query switches from writing to the adversary's space to writing to the oracle itself. This allows a reduction to simulate the oracle by simply recording information about the adversary's query in the Fourier domain. We then use this new technique to show the indifferentiability of the Merkle-Damgard domain extender for hash functions. We also give a proof of security for the Fujisaki-Okamoto transformation; previous proofs required modifying the scheme to include an additional hash term. Given the threat posed by quantum computers and the push toward quantum-resistant cryptosystems, our work represents an important tool for efficient post-quantum cryptosystems.

We provide a generic construction of non-interactive zero-knowledge (NIZK) schemes. Our construction is a refinement of Dwork and Naor’s (FOCS 2000) implementation of the hidden bits model using verifiable pseudorandom generators (VPRGs). Our refinement simplifies their construction and relaxes the necessary assumptions considerably. As a result of this conceptual improvement, we obtain interesting new instantiations: – A designated-verifier NIZK (with unbounded soundness) based on the computational Diffie-Hellman (CDH) problem. If a pairing is available, this NIZK becomes publicly verifiable. This constitutes the first fully secure CDH-based designated-verifier NIZKs (and more generally, the first fully secure designated-verifier NIZK from a non-generic assumption which does not already imply publicly-verifiable NIZKs), and it answers an open problem recently raised by Kim and Wu (CRYPTO 2018). – A NIZK based on the learning with errors (LWE) assumption, and assuming a non-interactive witness-indistinguishable (NIWI) proof system for bounded distance decoding (BDD). This simplifies and improves upon a recent NIZK from LWE that assumes a NIZK for BDD (Rothblum et al., PKC 2019).

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 $1/2-1/poly(n)$. 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 $1/2-1/poly(n)$ 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 $1/2-1/poly(n)$. 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 $\text{Search-BPP}^{SZK}$ (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 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 $\text{BPP}^\text{SZK})$.

We continue the study of statistical/computational tradeoffs in learning robust classifiers, following the recent work of Bubeck, Lee, Price and Razenshteyn who showed examples of classification tasks where (a) an efficient robust classifier exists, *in the small-perturbation regime*; (b) a non-robust classifier can be learned efficiently; but (c) it is computationally hard to learn a robust classifier, assuming the hardness of factoring large numbers. Indeed, the question of whether a robust classifier for their task exists in the large perturbation regime seems related to important open questions in computational number theory. In this work, we extend their work in three directions. First, we demonstrate classification tasks where computationally efficient robust classification is impossible, even when computationally unbounded robust classifiers exist. We rely on the hardness of decoding problems with preprocessing on codes and lattices. Second, we show hard-to-robustly-learn classification tasks *in the large-perturbation regime*. Namely, we show that even though an efficient classifier that is very robust (namely, tolerant to large perturbations) exists, it is computationally hard to learn any non-trivial robust classifier. Our first task relies on the existence of one-way functions, a minimal assumption in cryptography, and the second on the hardness of the learning parity with noise problem. In the latter setting, not only does a non-robust classifier exist, but also an efficient algorithm that generates fresh new labeled samples given access to polynomially many training examples (termed as generation by Kearns et. al. (1994)). Third, we show that any such counterexample implies the existence of cryptographic primitives such as one-way functions or even forms of public-key encryption. This leads us to a win-win scenario: either we can quickly learn an efficient robust classifier, or we can construct new instances of popular and useful cryptographic primitives.

In 2005, [2] Philippe Guillot presented a new construction of Boolean functions using linear codes as an extension of Maiorana-McFarland's construction of bent functions. In this paper, we study a new family of Boolean functions with cryptographically strong properties such as non- linearity, propagation criterion, resiliency and balance. The construction of cryptographically strong boolean functions is a daunting task and there is currently a wide range of algebraic techniques and heuristics for constructing such functions , however these methods can be complex, computationally difficult to implement and not always produce a sufficient variety of functions. We present in this paper a construction of Boolean functions using algebraic codes following Guillot's work.

We describe an algorithm to solve the approximate Shortest Vector Problem for lattices corresponding to ideals of the ring of integers of an arbitrary number field $K$. This algorithm has a pre-processing phase, whose run-time is exponential in $\log |\Delta|$ with $\Delta$ the discriminant of $K$. Importantly, this pre-processing phase depends only on $K$. The pre-processing phase outputs an advice, whose bit-size is no more than the run-time of the query phase. Given this advice, the query phase of the algorithm takes as input any ideal $I$ of the ring of integers, and outputs an element of $I$ which is at most $\exp(\widetilde O((\log |\Delta|)^{\alpha+1}/n))$ times longer than a shortest non-zero element of $I$ (with respect to the Euclidean norm of its canonical embedding). This query phase runs in time and space $\exp(\widetilde O( (\log |\Delta|)^{\max(2/3, 1-2\alpha)}))$ in the classical setting, and $\exp(\widetilde O((\log |\Delta|)^{1-2\alpha}))$ in the quantum setting. The parameter $\alpha$ can be chosen arbitrarily in $[0,1/2]$. Both correctness and cost analyses rely on heuristic assumptions, whose validity is consistent with experiments. The algorithm builds upon the algorithms from Cramer al. [EUROCRYPT 2016] and Cramer et al. [EUROCRYPT 2017]. It relies on the framework from Buchmann [Séminaire de théorie des nombres 1990], which allows to merge them and to extend their applicability from prime-power cyclotomic fields to all number fields. The cost improvements are obtained by allowing precomputations that depend on the field only.

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 $S,2^D$, 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 a quarter 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.

A $k$-collision for a compressing hash function $H$ is a set of $k$ distinct inputs that all map to the same output. In this work, we show that for any constant $k$, $\Theta\left(N^{\frac{1}{2}(1-\frac{1}{2^k-1})}\right)$ quantum queries are both necessary and sufficient to achieve a $k$-collision with constant probability. This improves on both the best prior upper bound (Hosoyamada et al., ASIACRYPT 2017) and provides the first non-trivial lower bound, completely resolving the problem.

In state-of-the-art design paradigm, time, space and power efficiency are considered the primary design constraints. Quite often, this approach adversely impacts the security of the overall system, especially when security is adopted as a countermeasure after some vulnerability is identified. In this position paper, we motivate the idea that security should also be considered as an architectural design constraint in addition to time, space and power. We show that security and efficiency objectives along the three design axes of time, space and power are in fact tightly coupled while identifying that security stands in direct contrast with them across all layers of architectural design. We attempt to prove our case utilizing a proof-by-evidence approach wherein we refer to various works across literature that explicitly imply the eternal conflict between security and efficiency. Thus, security has to be treated as a design constraint from the very beginning. Additionally, we advocate a security-aware design flow starting from the choice of cryptographic primitives, protocols and system design.

A game-based cryptographic proof is a relation that establishes equivalence between probabilistic sequences of actions by real and ideal world players. The author of a proof selects a hardness assumption system for their proof upon which to base their subsequent statements. In this paper, we prove the existence of proof-invariant transformations for varying hardness assumptions. We show that for two systems satisfying certain algebraic properties any proof in one system has an equivalent valid proof in the other. This validates Kurosawa’s remark about the existence of proof similarities. Our result implies a correspondence between the Learning With Errors (LWE) problems and both the Elliptic Curve Discrete Log problem (ECDLP) and the Discrete Logarithm (DLOG) problem. To illustrate this result, we provide a series of example transformations in the appendix. The concrete result of this paper is a prototype proof translation tool.

Property-preserving hashing is a method of compressing a large input x into a short hash h(x) in such a way that given h(x) and h(y), one can compute a property P(x, y) of the original inputs. The idea of property-preserving hash functions underlies sketching, compressed sensing and locality-sensitive hashing. Property-preserving hash functions are usually probabilistic: they use the random choice of a hash function from a family to achieve compression, and as a consequence, err on some inputs. Traditionally, the notion of correctness for these hash functions requires that for every two inputs x and y, the probability that h(x) and h(y) mislead us into a wrong prediction of P(x, y) is negligible. As observed in many recent works (incl. Mironov, Naor and Segev, STOC 2008; Hardt and Woodruff, STOC 2013; Naor and Yogev, CRYPTO 2015), such a correctness guarantee assumes that the adversary (who produces the offending inputs) has no information about the hash function, and is too weak in many scenarios. We initiate the study of adversarial robustness for property-preserving hash functions, provide definitions, derive broad lower bounds due to a simple connection with communication complexity, and show the necessity of computational assumptions to construct such functions. Our main positive results are two candidate constructions of property-preserving hash functions (achieving different parameters) for the (promise) gap-Hamming property which checks if x and y are “too far” or “too close”. Our first construction relies on generic collision-resistant hash functions, and our second on a variant of the syndrome decoding assumption on low-density parity check codes.

We study the isogeny graphs of supersingular elliptic curves over finite fields, with an emphasis on the vertices corresponding to elliptic curves of $j$-invariant 0 and 1728.

The two traditional hard problems underlying the security of lattice-based cryptography are the shortest vector problem (SVP) and the closest vector problem (CVP). For a long time, lattice enumeration was considered the fastest method for solving these problems in high dimensions, but recent work on memory-intensive methods has resulted in lattice sieving overtaking enumeration both in theory and in practice. Some of the recent improvements [Ducas, Eurocrypt 2018; Laarhoven-Mariano, PQCrypto 2018; Albrecht-Ducas-Herold-Kirshanova-Postlethwaite-Stevens, Eurocrypt 2019] are based on the fact that these methods find more than just one short lattice vector, and this additional data can be reused effectively later on to solve other, closely related problems faster. Similarly, results for the preprocessing version of CVP (CVPP) have demonstrated that once this initial data has been generated, instances of CVP can be solved faster than when solving them directly, albeit with worse memory complexities [Laarhoven, SAC 2016]. In this work we study CVPP in terms of approximate Voronoi cells, and obtain better time and space complexities using randomized slicing, which is similar in spirit to using randomized bases in lattice enumeration [Gama-Nguyen-Regev, Eurocrypt 2010]. With this approach, we improve upon the state-of-the-art complexities for CVPP, both theoretically and experimentally, with a practical speedup of several orders of magnitude compared to non-preprocessed SVP or CVP. Such a fast CVPP solver may give rise to faster enumeration methods, where the CVPP solver is used to replace the bottom part of the enumeration tree, consisting of a batch of CVP instances in the same lattice. Asymptotically, we further show that we can solve an exponential number of instances of CVP in a lattice in essentially the same amount of time and space as the fastest method for solving just one CVP instance. This is in line with various recent results, showing that perhaps the biggest strength of memory-intensive methods lies in being able to reuse the generated data several times. Similar to [Ducas, Eurocrypt 2018], this further means that we can achieve a ``few dimensions for free'' for sieving for SVP or CVP, by doing $\Theta(d/\log d)$ levels of enumeration on top of a CVPP solver based on approximate Voronoi cells.

Functional encryption (FE) is advanced encryption that enables us to issue functional decryption keys where functions are hardwired. When we decrypt a ciphertext of a message $m$ by a functional decryption key where a function $f$ is hardwired, we can obtain $f(m)$ and nothing else. We say FE is selectively or adaptively secure when target messages are chosen at the beginning or after function queries are sent, respectively. In the weakly-selective setting, function queries are also chosen at the beginning. We say FE is single-key/collusion-resistant when it is secure against adversaries that are given only-one/polynomially-many functional decryption keys, respectively. We say FE is sublinearly-succinct/succinct when the running time of an encryption algorithm is sublinear/poly-logarithmic in the function description size, respectively. In this study, we propose a generic transformation from weakly-selectively secure, single-key, and sublinearly-succinct (we call ``building block'') PKFE for circuits into adaptively secure, collusion-resistant, and succinct (we call ``fully-equipped'') one for circuits. We assume only the existence of the building block PKFE for circuits. That is, our transformation relies on neither concrete assumptions such as learning with errors nor indistinguishability obfuscation (IO). This is the first generic construction of fully-equipped PKFE that does not rely on IO. As side-benefits of our results, we obtain the following primitives from the building block PKFE for circuits: (1) laconic oblivious transfer (2) succinct garbling scheme for Turing machines (3) selectively secure, collusion-resistant, and succinct PKFE for Turing machines (4) low-overhead adaptively secure traitor tracing (5) key-dependent-message secure and leakage-resilient public-key encryption. We also obtain a generic transformation from simulation-based adaptively secure garbling schemes that satisfy a natural decomposability property into adaptively indistinguishable garbling schemes whose online complexity does not depend on the output length.

We revisit the concept of *non-malleable* secret sharing (Goyal and Kumar, STOC 2018) in the computational setting. In particular, under the assumption of one-to-one one-way functions, we exhibit a *computationally* private, *threshold* secret sharing scheme satisfying all of the following properties. -) Continuous non-malleability: No computationally-bounded adversary tampering independently with all the shares can produce mauled shares that reconstruct to a value related to the original secret. This holds even in case the adversary can tamper *continuously*, for an *unbounded* polynomial number of times, with the same target secret sharing, where the next sequence of tampering functions, as well as the subset of shares used for reconstruction, can be chosen *adaptively* based on the outcome of previous reconstructions. -) Resilience to noisy leakage: Non-malleability holds even if the adversary can additionally leak information independently from all the shares. There is no bound on the length of leaked information, as long as the overall leakage does not decrease the min-entropy of each share by too much. -) Improved rate: The information rate of our final scheme, defined as the ratio between the size of the message and the maximal size of a share, asymptotically approaches 1 when the message length goes to infinity. Previous constructions achieved information-theoretic security, sometimes even for arbitrary access structures, at the price of *at least one* of the following limitations: (i) Non-malleability only holds against one-time tampering attacks; (ii) Non-malleability holds against a bounded number of tampering attacks, but both the choice of the tampering functions and of the sets used for reconstruction is non-adaptive; (iii) Information rate asymptotically approaching zero; (iv) No security guarantee in the presence of leakage.

We observe that if a party breaks one cryptographic assumption, construction, or system, then it can reduce the trust in any other. This highlights a shortcoming in the common interpretation of the provable security paradigm that may lead to unwarranted trust. This may have practical implications. Then we argue that the provable security paradigm remains sound in applications provided that assumptions are made with care. We also strengthen the argument for the study of combiners and constructions based on generic assumptions, and transparent standardization processes in applied cryptography.

Non-signaling games are an important object of study in the theory of computation, for their role both in quantum information and in (classical) cryptography. In this work, we study the behavior of these games under parallel repetition. We show that, unlike the situation both for classical games and for two-player non-signaling games, there are $k$-player non-signaling games (for $k \ge 3$) whose values do not tend to $0$ with sufficient parallel repetition. In fact, parallel repetition sometimes does not decrease their value whatsoever. We show that in general: 1. Every game's non-signaling value under parallel repetition is either lower bounded by a positive constant or decreases exponentially with the number of repetitions. 2. Exponential decrease occurs if and only if the game's sub-non-signaling value (Lancien and Winter, CJTCS '16) is less than $1$.

Distributional collision resistance is a relaxation of collision resistance that only requires that it is hard to sample a collision (x,y) where x is uniformly random and y is uniformly random conditioned on colliding with x. The notion lies between one-wayness and collision resistance, but its exact power is still not well-understood. On one hand, distributional collision resistant hash functions cannot be built from one-way functions in a black-box way, which may suggest that they are stronger. On the other hand, so far, they have not yielded any applications beyond one-way functions. Assuming distributional collision resistant hash functions, we construct constant-round statistically hiding commitment scheme. Such commitments are not known based on one-way functions and are impossible to obtain from one-way functions in a black-box way. Our construction relies on the reduction from inaccessible entropy generators to statistically hiding commitments by Haitner et al. (STOC '09). In the converse direction, we show that two-message statistically hiding commitments imply distributional collision resistance, thereby establishing a loose equivalence between the two notions. A corollary of the first result is that constant-round statistically hiding commitments are implied by average-case hardness in the class SZK (which is known to imply distributional collision resistance). This implication seems to be folklore, but to the best of our knowledge has not been proven explicitly. We provide yet another proof of this implication, which is arguably more direct than the one going through distributional collision resistance.

Techniques for eliminating variables from a system of nonlinear equations are used to find solutions of the system. We discuss how these methods can be used to attack certain types of symmetric block ciphers, by solving sets of equations arising from known plain text attacks. The systems of equations corresponding to these block ciphers have the characteristics that the solution is determined by a small subset of the variables (i.e., the secret key), and also that it is known that there always exists at least one solution (again corresponding to the key which is actually used in the encryption). It turns out that some toy ciphers can be solved simpler than anticipated by this method, and that the method can take advantage of overdetermined systems.

Algebraic structure lies at the heart of much of Cryptomania as we know it. An interesting question is the following: instead of building (Cryptomania) primitives from concrete assumptions, can we build them from simple Minicrypt primitives endowed with additional algebraic structure? In this work, we affirmatively answer this question by adding algebraic structure to the following Minicrypt primitives: • One-Way Function (OWF) • Weak Unpredictable Function (wUF) • Weak Pseudorandom Function (wPRF) The algebraic structure that we consider is group homomorphism over the input/output spaces of these primitives. We also consider a “bounded” notion of homomorphism where the primitive only supports an a priori bounded number of homomorphic operations in order to capture lattice-based and other “noisy” assumptions. We show that these structured primitives can be used to construct many cryptographic protocols. In particular, we prove that: • (Bounded) Homomorphic OWFs (HOWFs) imply collision-resistant hash functions, Schnorr-style signatures, and chameleon hash functions. • (Bounded) Input-Homomorphic weak UFs (IHwUFs) imply CPA-secure PKE, non-interactive key exchange, trapdoor functions, blind batch encryption (which implies anonymous IBE, KDM-secure and leakage-resilient PKE), CCA2 deterministic PKE, and hinting PRGs (which in turn imply transformation of CPA to CCA security for ABE/1-sided PE). • (Bounded) Input-Homomorphic weak PRFs (IHwPRFs) imply PIR, lossy trapdoor functions, OT and MPC (in the plain model). In addition, we show how to realize any CDH/DDH-based protocol with certain properties in a generic manner using IHwUFs/IHwPRFs, and how to instantiate such a protocol from many concrete assumptions. We also consider primitives with substantially richer structure, namely Ring IHwPRFs and L-composable IHwPRFs. In particular, we show the following: • Ring IHwPRFs with certain properties imply FHE. • 2-composable IHwPRFs imply (black-box) IBE, and $L$-composable IHwPRFs imply non-interactive $(L + 1)$-party key exchange. Our framework allows us to categorize many cryptographic protocols based on which structured Minicrypt primitive implies them. In addition, it potentially makes showing the existence of many cryptosystems from novel assumptions substantially easier in the future.

Recently, Faust et al. (TCC'14) introduced the notion of continuous non-malleable codes (CNMC), which provides stronger security guarantees than standard non-malleable codes, by allowing an adversary to tamper with the codeword in continuous way instead of one-time tampering. They also showed that CNMC with information theoretic security cannot be constructed in 2-split-state tampering model, and presented a construction of the same in CRS (common reference string) model using collision-resistant hash functions and non-interactive zero-knowledge proofs. In this work, we ask if it is possible to construct CNMC from weaker assumptions. We answer this question by presenting lower as well as upper bounds. Specifically, we show that it is impossible to construct 2-split-state CNMC, with no CRS, for one-bit messages from any falsifiable assumption, thus establishing the lower bound. We additionally provide an upper bound by constructing 2-split-state CNMC for one-bit messages, assuming only the existence of a family of injective one way functions. We also present a construction of 4-split-state CNMC for multi-bit messages in CRS model from the same assumptions. Additionally, we present definitions of the following new primitives: (1) One-to-one commitments, and (2) Continuous Non-Malleable Randomness Encoders, which may be of independent interest.

We study the *rate* of so-called *continuously* non-malleable codes, which allow to encode a message in such a way that (possibly adaptive) continuous tampering attacks on the codeword yield a decoded value that is unrelated to the original message. Our results are as follows: -) For the case of bit-wise independent tampering, we establish the existence of rate-one continuously non-malleable codes with information-theoretic security, in the plain model. -) For the case of split-state tampering, we establish the existence of rate-one continuously non-malleable codes with computational security, in the (non-programmable) random oracle model. We further exhibit a rate-1/2 code and a rate-one code in the common reference string model, but the latter only withstands *non-adaptive* tampering. It is well known that computational security is inherent for achieving continuous non-malleability in the split-state model (even in the presence of non-adaptive tampering). Continuously non-malleable codes are useful for protecting *arbitrary* cryptographic primitives against related-key attacks, as well as for constructing non-malleable public-key encryption schemes. Our results directly improve the efficiency of these applications.

Many efficient data structures use randomness, allowing them to improve upon deterministic ones. Usually, their efficiency and correctness are analyzed using probabilistic tools under the assumption that the inputs and queries are independent of the internal randomness of the data structure. In this work, we consider data structures in a more robust model, which we call the adversarial model. Roughly speaking, this model allows an adversary to choose inputs and queries adaptively according to previous responses. Specifically, we consider a data structure known as "Bloom filter" and prove a tight connection between Bloom filters in this model and cryptography. A Bloom filter represents a set $S$ of elements approximately, by using fewer bits than a precise representation. The price for succinctness is allowing some errors: for any $x \in S$ it should always answer `Yes', and for any $x \notin S$ it should answer `Yes' only with small probability. In the adversarial model, we consider both efficient adversaries (that run in polynomial time) and computationally unbounded adversaries that are only bounded in the number of queries they can make. For computationally bounded adversaries, we show that non-trivial (memory-wise) Bloom filters exist if and only if one-way functions exist. For unbounded adversaries we show that there exists a Bloom filter for sets of size $n$ and error $\varepsilon$, that is secure against $t$ queries and uses only $O(n \log{\frac{1}{\varepsilon}}+t)$ bits of memory. In comparison, $n\log{\frac{1}{\varepsilon}}$ is the best possible under a non-adaptive adversary.

A cryptographic function with a fixed-length output, such as a block cipher, hash function, or message authentication code (MAC), should behave as a random mapping. The mapping's randomness can be evaluated with statistical tests. Statistical test suites typically used to evaluate cryptographic functions, such as the NIST test suite, are not well-suited for testing fixed-output-length cryptographic functions. Also, these test suites employ a frequentist approach, making it difficult to obtain an overall evaluation of the mapping's randomness. This paper describes CryptoStat, a test suite that overcomes the aforementioned deficiencies. CryptoStat is specifically designed to test the mappings of fixed-output-length cryptographic functions, and CryptoStat employs a Bayesian approach that quite naturally yields an overall evaluation of the mappings' randomness. Results of applying CryptoStat to reduced-round and full-round versions of the AES block ciphers and the SHA-1 and SHA-2 hash functions are reported; the results are analyzed to determine the algorithms' randomness margins.

The current paper improves the number of queries of the previous quantum multi-collision finding algorithms presented by Hosoyamada et al. at Asiacrypt 2017. Let an $l$-collision be a tuple of $l$ distinct inputs that result in the same output of a target function. In cryptology, it is important to study how many queries are required to find $l$-collisions for random functions of which domains are larger than ranges. The previous algorithm finds an $l$-collision for a random function by recursively calling the algorithm for finding $(l-1)$-collisions, and it achieves the average quantum query complexity of $O(N^{(3^{l-1}-1) / (2 \cdot 3^{l-1})})$, where $N$ is the range size of target functions. The new algorithm removes the redundancy of the previous recursive algorithm so that different recursive calls can share a part of computations. The new algorithm finds an $l$-collision for random functions with the average quantum query complexity of $O(N^{(2^{l-1}-1) / (2^{l}-1)})$, which improves the previous bound for all $l\ge 3$ (the new and previous algorithms achieve the optimal bound for $l=2$). More generally, the new algorithm achieves the average quantum query complexity of $O\left(c^{3/2}_N N^{\frac{2^{l-1}-1}{ 2^{l}-1}}\right)$ for a random function $f\colon X\to Y$ such that $|X| \geq l \cdot |Y| / c_N$ for any $1\le c_N \in o(N^{\frac{1}{2^l - 1}})$. With the same query complexity, it also finds a multiclaw for random functions, which is harder to find than a multicollision.

In this paper, we give a no-signaling linear probabilistically checkable proof (PCP) system for polynomial-time deterministic computation, i.e., a PCP system for P such that (1) the honest PCP oracle is a linear function and (2) the soundness holds against any (computational) no-signaling cheating prover, who is allowed to answer each query according to a distribution that depends on the entire query set in a certain way. To the best of our knowledge, our construction is the first PCP system that satisfies these two properties simultaneously. As an application of our PCP system, we obtain a 2-message delegating computation scheme by using a known transformation. Compared with the existing 2-message delegating computation schemes that are based on standard cryptographic assumptions, our scheme requires preprocessing but has a simpler structure and makes use of different (possibly cheaper) standard cryptographic primitives, namely additive/multiplicative homomorphic encryption schemes.

Non-malleable asymmetric encryption schemes which prove plaintext knowledge are sufficient for secrecy in some domains. For example, ballot secrecy in voting. In these domains, some applications derive encryption schemes by coupling malleable ciphertexts with proofs of plaintext knowledge, without evidence that the sufficient condition (for secrecy) is satisfied nor an independent security proof (of secrecy). Consequently, it is unknown whether these applications satisfy desirable secrecy properties. In this article, we propose a generic construction for such a coupling and show that our construction produces non-malleable encryption schemes which prove plaintext knowledge. Furthermore, we show how our results can be used to prove ballot secrecy of voting systems. Accordingly, we facilitate the development of applications satisfying their security objectives.

Robustly reusable Fuzzy Extractor (rrFE) considers reusability and robustness simultaneously. We present two approaches to the generic construction of rrFE. Both of approaches make use of a secure sketch and universal hash functions. The first approach also employs a special pseudo-random function (PRF), namely unique-input key-shift (ui-ks) secure PRF, and the second uses a key-shift secure auxiliary-input authenticated encryption (AIAE). The ui-ks security of PRF (resp. key-shift security of AIAE), together with the homomorphic properties of secure sketch and universal hash function, guarantees the reusability and robustness of rrFE. Meanwhile, we show two instantiations of the two approaches respectively. The first instantiation results in the first rrFE from the LWE assumption, while the second instantiation results in the first rrFE from the DDH assumption over non-pairing groups.

We investigate the differential properties of a construction in which a given function $F : \mathbb{F}_{2^n} \rightarrow \mathbb{F}_{2^n}$ is modified at $K \in \mathbb{N}$ points in order to obtain a new function $G$. This is motivated by the question of determining the minimum Hamming distance between two APN functions and can be seen as a generalization of a previously studied construction in which a given function is modified at a single point. We derive necessary and sufficient conditions which the derivatives of $F$ must satisfy for $G$ to be APN, and use these conditions as the basis for an efficient filtering procedure for searching for APN functions whose value differs from that of a given APN function $F$ at a given set of points. We define a quantity $m_F$ related to $F$ counting the number of derivatives of a given type, and derive a lower bound on the distance between an APN function $F$ and its closest APN neighbor in terms of $m_F$. Furthermore, the value $m_F$ is shown to be invariant under CCZ-equivalence and easier to compute in the case of quadratic functions. We give a formula for $m_F$ in the case of $F(x) = x^3$ which allows us to express a lower bound on the distance between $F(x)$ and the closest APN function in terms of the dimension $n$ of the underlying field. We observe that this distance tends to infinity with $n$. We also compute $m_F$ and the distance to the closest APN function for a representative $F$ from each of the switching classes over $\mathbb{F}_{2^n}$ for $4 \le n \le 8$. For a given function $F$ and value $v$, we describe an efficient method for finding all sets of points $\{ u_1, u_2, \dots, u_K \}$ such that setting $G(u_i) = F(u_i) + v$ and $G(x) = F(x)$ for $x \ne u_i$ is APN.

In the area of distributed graph algorithms a number of network's entities with local views solve some computational task by exchanging messages with their neighbors. Quite unfortunately, an inherent property of most existing distributed algorithms is that throughout the course of their execution, the nodes get to learn not only their own output but rather learn quite a lot on the inputs or outputs of many other entities. This leakage of information might be a major obstacle in settings where the output (or input) of network's individual is a private information (e.g., distributed networks of selfish agents, decentralized digital currency such as Bitcoin). While being quite an unfamiliar notion in the classical distributed setting, the notion of secure multi-party computation (MPC) is one of the main themes in the Cryptographic community. The existing secure MPC protocols do not quite fit the framework of classical distributed models in which only messages of bounded size are sent on graph edges in each round. In this paper, we introduce a new framework for \emph{secure distributed graph algorithms} and provide the first \emph{general compiler} that takes any "natural" non-secure distributed algorithm that runs in $r$ rounds, and turns it into a secure algorithm that runs in $\widetilde{O}(r \cdot D \cdot poly(\Delta))$ rounds where $\Delta$ is the maximum degree in the graph and $D$ is its diameter. A "natural" distributed algorithm is one where the local computation at each node can be performed in polynomial time. An interesting advantage of our approach is that it allows one to decouple between the price of locality and the price of \emph{security} of a given graph function $f$. The security of the compiled algorithm is information-theoretic but holds only against a semi-honest adversary that controls a single node in the network. This compiler is made possible due to a new combinatorial structure called \emph{private neighborhood trees}: a collection of $n$ trees $T(u_1),\ldots, T(u_n)$, one for each vertex $u_i \in V(G)$, such that each tree $T(u_i)$ spans the neighbors of $u_i$ {\em without going through $u_i$}. Intuitively, each tree $T(u_i)$ allows all neighbors of $u_i$ to exchange a \emph{secret} that is hidden from $u_i$, which is the basic graph infrastructure of the compiler. In a $(d,c)$-private neighborhood trees each tree $T(u_i)$ has depth at most $d$ and each edge $e \in G$ appears in at most $c$ different trees. We show a construction of private neighborhood trees with $d=\widetilde{O}(\Delta \cdot D)$ and $c=\widetilde{O}(D)$, both these bounds are \emph{existentially} optimal.

The universal composability paradigm allows for the modular design and analysis of cryptographic protocols. It has been widely and successfully used in cryptography. However, devising a coherent yet simple and expressive model for universal composability is, as the history of such models shows, highly non-trivial. For example, several partly severe problems have been pointed out in the literature for the UC model. In this work, we propose a coherent model for universal composability, called the IITM model (``Inexhaustible Interactive Turing Machine''). A main feature of the model is that it is stated without a priori fixing irrelevant details, such as a specific way of addressing of machines by session and party identifiers, a specific modeling of corruption, or a specific protocol hierarchy. In addition, we employ a very general notion of runtime. All reasonable protocols and ideal functionalities should be expressible based on this notion in a direct and natural way, and without tweaks, such as (artificial) padding of messages or (artificially) adding extra messages. Not least because of these features, the model is simple and expressive. Also the general results that we prove, such as composition theorems, hold independently of how such details are fixed for concrete applications. Being inspired by other models for universal composability, in particular the UC model and because of the flexibility and expressivity of the IITM model, conceptually, results formulated in these models directly carry over to the IITM model.

In this work, we consider the natural goal of designing secret sharing schemes that ensure security against a powerful adaptive adversary who may learn some ``leaked'' information about all the shares. We say that a secret sharing scheme is $p$-party leakage-resilient, if the secret remains statistically hidden even after an adversary learns a bounded amount of leakage, where each bit of leakage can depend jointly on the shares of an adaptively chosen subset of $p$ parties. A lot of works have focused on designing secret sharing schemes that handle individual and (mostly) non-adaptive leakage for (some) threshold secret sharing schemes [DP07,DDV10,LL12,ADKO15,GK18,BDIR18]. We give an unconditional compiler that transforms any standard secret sharing scheme with arbitrary access structure into a $p$-party leakage-resilient one for $p$ logarithmic in the number of parties. This yields the first secret sharing schemes secure against adaptive and joint leakage for more than two parties. As a natural extension, we initiate the study of leakage-resilient non-malleable secret sharing} and build such schemes for general access structures. We empower the computationally unbounded adversary to adaptively leak from the shares and then use the leakage to tamper with each of the shares arbitrarily and independently. Leveraging our $p$-party leakage-resilient schemes, we also construct such non-malleable secret sharing schemes: any such tampering either preserves the secret or completely `destroys' it. This improves upon the non-malleable secret sharing scheme of Goyal and Kumar (CRYPTO 2018) where no leakage was permitted. Leakage-resilient non-malleable codes can be seen as 2-out-of-2 schemes satisfying our guarantee and have already found several applications in cryptography [LL12,ADKO15,GKPRS18,GK18,CL18,OPVV18]. Our constructions rely on a clean connection we draw to communication complexity in the well-studied number-on-forehead (NOF) model and rely on functions that have strong communication-complexity lower bounds in the NOF model (in a black-box way). We get efficient $p$-party leakage-resilient schemes for $p$ upto $O(\log n)$ as our share sizes have exponential dependence on $p$. We observe that improving this dependence from $2^{O(p)}$ to $2^{o(p)}$ will lead to progress on longstanding open problems in complexity theory.

Lossy trapdoor functions (Peikert and Waters, STOC 2008 and SIAM J. Computing 2011) imply, via black-box transformations, a number of interesting cryptographic primitives, including chosen-ciphertext secure public-key encryption. Kiltz, O'Neill, and Smith (CRYPTO 2010) showed that the RSA trapdoor permutation is lossy under the Phi-hiding assumption, but syntactically it is not a lossy trapdoor function since it acts on Z_N and not on strings. Using a domain extension technique by Freeman et al. (PKC 2010 and J. Cryptology 2013) it can be extended to a lossy trapdoor permutation, but with considerably reduced lossiness. In this work we give new constructions of lossy trapdoor permutations from the Phi-hiding assumption, the quadratic residuosity assumption, and the decisional composite residuosity assumption, all with improved lossiness. Furthermore, we propose the first all-but-one lossy trapdoor permutation from the Phi-hiding assumption. A technical vehicle used for achieving this is a novel transform that converts trapdoor functions with index-dependent domain into trapdoor functions with fixed domain.

Maximally nonlinear Boolean functions in $n$ variables, where n is even, are called bent functions. There are several ways to represent Boolean functions. One of the most useful is via algebraic normal form (ANF). What can we say about ANF of a bent function? We try to collect all known and new facts related to ANF of a bent function. A new problem in bent functions is stated and studied: is it true that a linear, quadratic, cubic, etc. part of ANF of a bent function can be arbitrary? The case of linear part is well studied before. In this paper we prove that a quadratic part of a bent function can be arbitrary too.

We generalize and abstract the problem of extracting a witness from a prover of a special sound protocol into a combinatorial problem induced by a sequence of matroids and a predicate, and present a parametrized algorithm for solving this problem. The parametrization provides a tight tradeoff between the running time and the extraction error of the algorithm, which allows optimizing the parameters to minimize: the soundness error for interactive proofs, or the extraction time for proofs of knowledge. In contrast to previous work we bound the distribution of the running time and not only the expected running time. Tail bounds give a tighter analysis when applied recursively and concentrated running time.

We construct a genus 2 curve inside the product of 2 elliptic curves. The formula for this construction has appeared in a previous paper. The current paper discusses how this formula arises naturally by using some theory of elliptic Kummer surfaces

We propose a new method for reducing complexity of the pairing comparisons based on polynomials. Thought the construction introduces uncertainty into (usually deterministic) checks, it is easily quantifiable and in most cases extremely small. The application to CL-LRSW signature verification under n messages and group order q allows to reduce the number of computed pairings from 4n down to just 4, while the introduced uncertainty is just (2n-1)/q.

Differing-inputs obfuscation (diO), first proposed by Barak et. al. [4], provides stronger security than that provided by indistinguishability obfuscation (iO). An iO scheme provides indistinguishability between the obfuscations of two programs that are equivalent and have the same length of description. A diO scheme ensures that the obfuscations of two efficiently generated programs with the same description length are indistinguishable if it is hard to find an input on which their outputs differ. Ananth et. al. [1] showed the definition of diO with respect to arbitrary auxiliary inputs. However, Garg et al. [19] showed that the existence of this kind of diO contradicts a certain “special-purpose obfuscation” conjecture. Ishai, Pandey and Sahai [23] suggested a diO variant called public-coin diO, which requires the auxiliary input to be a public random string and given as input to all relevant algorithms. They gave a construction of public-coin diO by assuming the existence of public-coin differing-inputs obfuscator for NC^1 circuits. In this paper, we use a slightly different definition, called public-coin-dependent diO. It allows the obfuscation algorithm to additionally take as input the random coins used to sample the circuit pair (including the circuit to be obfuscated) and thus the obfuscation algorithm can use the property of the circuit pair. We first construct a public-coin differing-inputs obfuscator for a class of new defined function with iO and point obfuscation with auxiliary input (AIPO). And then we use it to complete the public-coin-dependent diO for any pair of circuits that are hard to be found an input on which their outputs differ. The constructions are based on secure iO schemes for NC^1, fully homomorphic encryption scheme, and the existence of AIPO. Besides, we show the applications of our constructions.

Let $\mathbb G_n$ be the subgroup of elements of odd order in the group $\mathbb Z_n^\star$ and let $\mathcal U(\mathbb G_n)$ be the uniform probability distribution on $\mathbb G_n$. In this paper, we establish a probabilistic polynomial-time reduction from finding a nontrivial divisor of a composite number $n$ to finding a nontrivial relation between $l$ elements chosen independently and uniformly at random from $\mathbb G_n$, where $l\ge1$ is given in unary as a part of the input. Assume that finding a nontrivial divisor of a random number in some set $N$ of composite numbers (for a given security parameter) is a computationally hard problem. Then, using the above-mentioned reduction, we prove that the family $\{(\mathbb G_n,\mathcal U(\mathbb G_n))\}_{n\in N}$ of computational abelian groups is weakly pseudo-free. The disadvantage of this result is that the probability ensemble $\{\mathcal U(\mathbb G_n)\}_{n\in N}$ is not polynomial-time samplable. To overcome this disadvantage, we construct a polynomial-time computable function $\nu\colon D\to N$ (where $D\subseteq\{0,1\}^*$) and a polynomial-time samplable probability ensemble $\{\mathcal G_d\}_{d\in D}$ (where $\mathcal G_d$ is a distribution on $\mathbb G_{\nu(d)}$ for each $d\in D$) such that the family $\{(\mathbb G_{\nu(d)},\mathcal G_d)\}_{d\in D}$ of computational abelian groups is weakly pseudo-free.

This technical report discusses three subtleties related to the widely used notion of differential privacy (DP). First, we discuss how the choice of a distinguisher influences the privacy notion and why we should always have a distinguisher if we consider approximate DP. Secondly, we draw a line between the very intuitive probabilistic differential privacy (with probability $1-\delta$ we have $\varepsilon$-DP) and the commonly used approximate differential privacy ($(\varepsilon,\delta)$-DP). Finally we see that and why probabilistic differential privacy (and similar notions) are not complete under post-processing, which has significant implications for notions used in the literature.

It is well-known that non-comparison-based techniques can allow us to sort $n$ elements in $o(n \log n)$ time on a Random-Access Machine (RAM). On the other hand, it is a long-standing open question whether (non-comparison-based) circuits can sort $n$ elements from the domain $[1..2^k]$ with $o(k n \log n)$ boolean gates. We consider weakened forms of this question: first, we consider a restricted class of sorting where the number of distinct keys is much smaller than the input length; and second, we explore Oblivious RAMs and probabilistic circuit families, i.e., computational models that are somewhat more powerful than circuits but much weaker than RAM. We show that Oblivious RAMs and probabilistic circuit families can sort $o(\log n)$-bit keys in $o(n \log n)$ time or $o(k n \log n)$ circuit complexity where $n$ is the input length. Our algorithms work in the balls-and-bins model, i.e., not only can they sort an array of numerical keys --- if each key additionally carries an opaque ball, our algorithms can also move the balls into the correct order. We further show that in such a balls-and-bins model, it is impossible to sort $\Omega(\log n)$-bit keys in $o(n \log n)$ time, and thus the $o(\log n)$-bit-key assumption is necessary for overcoming the $n \log n$ barrier. Finally, we optimize the IO efficiency of our oblivious algorithms for RAMs --- we show that even the $1$-bit special case of our algorithm can solve open questions regarding whether there exist oblivious algorithms for tight compaction and selection in linear IO.

We construct efficient non-malleable codes (NMC) that are (computationally) secure against tampering by functions computable in any fixed polynomial time. Our construction is in the plain (no-CRS) model and requires the assumptions that (1) $\mathbf{E}$ is hard for $\mathbf{NP}$ circuits of some exponential $2^{\beta n}$ ($\beta>0$) size (widely used in the derandomization literature), (2) sub-exponential trapdoor permutations exist, and (3) $\mathbf{P}$ certificates with sub-exponential soundness exist. While it is impossible to construct NMC secure against arbitrary polynomial-time tampering (Dziembowski, Pietrzak, Wichs, ICS '10), the existence of NMC secure against $O(n^c)$-time tampering functions (for any fixed $c$), was shown (Cheraghchi and Guruswami, ITCS '14) via a probabilistic construction. An explicit construction was given (Faust, Mukherjee, Venturi, Wichs, Eurocrypt '14) assuming an untamperable CRS with length longer than the runtime of the tampering function. In this work, we show that under computational assumptions, we can bypass these limitations. Specifically, under the assumptions listed above, we obtain non-malleable codes in the plain model against $O(n^c)$-time tampering functions (for any fixed $c$), with codeword length independent of the tampering time bound. Our new construction of NMC draws a connection with non-interactive non-malleable commitments. In fact, we show that in the NMC setting, it suffices to have a much weaker notion called quasi non-malleable commitments---these are non-interactive, non-malleable commitments in the plain model, in which the adversary runs in $O(n^c)$-time, whereas the honest parties may run in longer (polynomial) time. We then construct a 4-tag quasi non-malleable commitment from any sub-exponential OWF and the assumption that $\mathbf{E}$ is hard for some exponential size $\mathbf{NP}$-circuits, and use tag amplification techniques to support an exponential number of tags.

Constrained pseudorandom functions (CPRFs) are a type of PRFs that allows one to derive a constrained key $\mathsf{K}_C$ from the master key $\mathsf{K}$. While the master key $\mathsf{K}$ allows one to evaluate on any input as a standard PRF, the constrained key $\mathsf{K}_C$ only allows one to evaluate on inputs $x$ such that $C(x) = 1$. Since the introduction of CPRFs by Boneh and Waters (ASIACRYPT'13), Kiayias et al. (CCS'13), and Boyle et al. (PKC'14), there have been various constructions of CPRFs. However, thus far, almost all constructions (from standard assumptions and non-trivial constraints) are only proven to be secure if at most one constrained key $\mathsf{K}_C$ is known to the adversary, excluding the very recent work of Davidson and Nishimaki (EPRINT'18). Considering the interesting applications of CPRFs such as ID-based non-interactive key exchange, we desire CPRFs that are collusion resistance with respect to the constrained keys. In this work, we make progress in this direction and construct a CPRF for the bit-fixing predicates that are collusion resistance for a constant number of constrained keys. Surprisingly, compared to the heavy machinery that was used by previous CPRF constructions, our construction only relies on the existence of one-way functions.

Constrained pseudorandom functions (CPRFs) allow learning modified PRF keys that can evaluate the PRF on a subset of the input space, or based on some sort of predicate. First introduced by Boneh and Waters [Asiacrypt 2013], they have been shown to be a useful cryptographic primitive with many applications. The full security definition of CPRFs requires the adversary to learn multiple constrained keys, a requirement for all of these applications. Unfortunately, existing constructions of CPRFs satisfying this security notion are only known from exceptionally strong cryptographic assumptions, such as indistinguishability obfuscation and the existence of multilinear maps, even for very weak predicates. CPRFs from more standard assumptions only satisfy security when one key is learnt. In this work, we give the first construction of a CPRF that can issue a constant number of constrained keys for bit-fixing predicates, from learning with errors (LWE). It also satisfies \(1\)-key privacy (otherwise known as constraint-hiding). Finally, our construction achieves fully adaptive security with polynomial security loss; the only construction to achieve such security under a standard assumption. Our technique represents a noted departure existing for CPRF constructions. We hope that it may lead to future constructions that can expose a greater number of keys, or consider more expressive predicates (such as circuit-based constraints).

The recent progress in key derivation (Barak at al. CRYPTO'11, Dodis Yu TCC'2013) introduced the concept of constrained profiles for attackers advantage, recognizing that security bounds can be significantly improved (alternatively: lots of randomness can be saved) when the advantage, as the function of the key, is bounded in mean or variance. This paper studies \emph{minimal requirements for keys} to achieve security under such restricted attackers. We frame the problem as characterizing \emph{pseudorandomness against constrained distinguishers} and show that minimal assumptions are respectively (a) high smooth min-entropy and (b) high smooth collision entropy. This matches the (folklore extension of) assumptions of previous works. Besides providing lower bounds, we offer more insights into this key derivation problem and elegant proof techniques of geometric flavor.

We study the problem of encrypting and authenticating quantum data in the presence of adversaries making adaptive chosen plaintext and chosen ciphertext queries. Classically, security games use string copying and comparison to detect adversarial cheating in such scenarios. Quantumly, this approach would violate no-cloning. We develop new techniques to overcome this problem: we use entanglement to detect cheating, and rely on recent results for characterizing quantum encryption schemes. We give denitions for (i.) ciphertext unforgeability , (ii.) indistinguishability under adaptive chosen-ciphertext attack, and (iii.) authenticated encryption. The restriction of each denition to the classical setting is at least as strong as the corresponding classical notion: (i) implies INT-CTXT, (ii) implies IND-CCA2, and (iii) implies AE. All of our new notions also imply QIND-CPA privacy. Combining one-time authentication and classical pseudorandomness, we construct schemes for each of these new quantum security notions, and provide several separation examples. Along the way, we also give a new denition of one-time quantum authentication which, unlike all previous approaches, authenticates ciphertexts rather than plaintexts.

We introduce Pseudo Flawed-smudging Generators (PFGs). A PFG is an expanding function whose outputs $\mathbf Y$ satisfy a weak form of pseudo-randomness. Roughly speaking, for some polynomial bound $B$, and every distribution $\chi$ over $B$-bounded noise vectors, it guarantees that the distribution of $(\mathbf e,\ \mathbf Y + \mathbf e)$ is indistinguishable from that of $(\mathbf e', \mathbf Y + \mathbf e)$, where $\mathbf e \gets \chi$ is a random sample from $\chi$, and $\mathbf e'$ is another independent sample from $\chi$ conditioned on agreeing with $\mathbf e$ at a few, $o(\lambda)$, coordinates. In other words, $\mathbf Y$ "hides" $\mathbf e$ at all but a few coordinates. We show that assuming LWE and the existence of constant-locality Pseudo-Random Generators (PRGs), there is a construction of IO from 1) a PFG that has polynomial stretch and polynomially bounded outputs, and 2) a Functional Encryption (FE) scheme able to compute this PFG. Such FE can be built from degree $d$ multilinear map if the PFG is computable by a degree $d$ polynomial. Toward basing IO on bilinear maps, inspired by [Ananth et. al. Eprint 2018], we further consider PFGs with partial pubic input --- they have the form $g(\mathbf{x}, \mathbf{y})$ and satisfy the aforementioned pseudo flawed-smudging property even when $\mathbf{x}$ is public. When using such PFGs, it suffices to replace FE with a weaker notion of partially hiding FE (PHFE) whose decryption reveals the public input $\mathbf{x}$ in addition to the output of the computation. We construct PHFE for polynomials $g$ that are quadratic in the private input $\mathbf{y}$, but have up to polynomial degree in the public input $\mathbf{x}$, subject to certain size constraints, from the SXDH assumption over bilinear map groups. Regarding candidates of PFGs with partial public input, we note that the family of cubic polynomials proposed by Ananth et. al. can serve as candidate PFGs, and can be evaluated by our PHFE from bilinear maps. Toward having more candidates, we present a transformation for converting the private input $\mathbf{x}$ of a constant-degree PFG $g(\mathbf{x}, \mathbf{y})$ into a public input, by hiding $\mathbf{x}$ as noises in LWE samples, provided that $\mathbf{x}$ is sampled from a LWE noise distribution and $g$ satisfies a stronger security property.

This tutorial demonstrates how cryptographic security notions, constructions, and game-based security proofs can be formalized using the CryptHOL framework. As a running example, we formalize a variant of the hash-based ElGamal encryption scheme and its IND-CPA security in the random oracle model. This tutorial assumes familiarity with Isabelle/HOL basics and standard cryptographic terminology.

Today’s blockchains do not scale in a meaningful sense. As more nodes join the system, the efficiency of the system (computation, communication, and storage) degrades, or at best stays constant. A leading idea for enabling blockchains to scale efficiency is the notion of sharding: different subsets of nodes handle different portions of the blockchain, thereby reducing the load for each individual node. However, existing sharding proposals achieve efficiency scaling by compromising on trust - corrupting the nodes in a given shard will lead to the permanent loss of the corresponding portion of data. We observe that sharding is similar to replication coding, which is known to be inefficient and fragile in the coding theory community. In this paper, we demonstrate a new protocol for coded storage and computation in blockchains. In particular, we propose PolyShard: “polynomially coded sharding” scheme that achieves information-theoretic upper bounds on the efficiency of the storage, system throughput, as well as on trust, thus enabling a truly scalable system. We provide simulation results that numerically demonstrate the performance improvement over state of the arts, and the scalability of the PolyShard system. Finally, we discuss potential enhancements, and highlight practical considerations in building such a system.

Motivatedbytheoreticalandpracticalconsiderations,anim- portant line of research is to design secure computation protocols that only make black-box use of cryptography. An important component in nearly all the black-box secure computation constructions is a black- box commit-and-prove protocol. A commit-and-prove protocol allows a prover to commit to a value and prove a statement about this value while guaranteeing that the committed value remains hidden. A black- box commit-and-prove protocol implements this functionality while only making black-box use of cryptography. In this paper, we build several tools that enable constructions of round- optimal, black-box commit and prove protocols. In particular, assuming injective one-way functions, we design the first round-optimal, black- box commit-and-prove arguments of knowledge satisfying strong privacy against malicious verifiers, namely: – Zero-knowledge in four rounds and, – Witness indistinguishability in three rounds. Prior to our work, the best known black-box protocols achieving commit- and-prove required more rounds. We additionally ensure that our protocols can be used, if needed, in the delayed-input setting, where the statement to be proven is decided only towards the end of the interaction. We also observe simple applications of our protocols towards achieving black-box four-round constructions of extractable and equivocal commitments. We believe that our protocols will provide a useful tool enabling several new constructions and easy round-efficient conversions from non-black- box to black-box protocols in the future.

We give a simple construction of indistinguishability obfus- cation for Turing machines where the time to obfuscate grows only with the description size of the machine and otherwise, independent of the running time and the space used. While this result is already known [Koppula, Lewko, and Waters, STOC 2015] from iO for circuits and injective pseudorandom generators, our construction and its analysis are conceptually much simpler. In particular, the main technical com- ponent in the proof of our construction is a simple combinatorial peb- bling argument [Garg and Srinivasan, EUROCRYPT 2018]. Our con- struction makes use of indistinguishability obfuscation for circuits and somewhere statistically binding hash functions.

The modeling of trapdoor permutations has evolved over the years. Indeed, finding an appropriate abstraction that bridges between the existing candidate constructions and the needs of applications has proved to be challenging. In particular, the notions of certifying permutations (Bellare and Yung, 96), enhanced and doubly enhanced trapdoor permutations (Goldreich, 04, 08, 11, Goldreich and Rothblum, 13) were added to bridge the gap between the modeling of trapdoor permutations and needs of applications. We identify an additional gap in the current abstraction of trapdoor permutations: Previous works implicitly assumed that it is easy to recognize elements in the domain, as well as uniformly sample from it, even for illegitimate function indices. We demonstrate this gap by using the (Bitansky-Paneth-Wichs, 16) doubly-enhanced trapdoor permutation family to instantiate the Feige-Lapidot-Shamir (FLS) paradigm for constructing non-interactive zero-knowledge (NIZK) protocols, and show that the resulting proof system is unsound. To close the gap, we propose a general notion of certifiably injective doubly enhanced trapdoor functions (DECITDFs), which provides a way of certifying that a given key defines an injective function over the domain defined by it, even when that domain is not efficiently recognizable and sampleable. We show that DECITDFs suffice for instantiating the FLS paradigm; more generally, we argue that certifiable injectivity is needed whenever the generation process of the function is not trusted. We then show two very different ways to construct DECITDFs: One is via the traditional method of RSA/Rabin with the Bellare-Yung certification mechanism, and the other using indistinguishability obfuscation and injective pseudorandom generators. In particular the latter is the first candidate injective trapdoor function, from assumptions other than factoring, that suffices for the FLS paradigm. Finally we observe that a similar gap appears also in other paths proposed in the literature for instantiating the FLS paradigm, specifically via verifiable pseudorandom generators and verifiable pseudorandom functions. Closing the gap there can be done in similar ways to the ones proposed here.

We consider the question of whether the security of unique digital signature schemes can be based on game-based cryptographic assumptions using linear-preserving black-box security reductions—that is, black-box reductions for which the security loss (i.e., the ratio between "work" of the adversary and the "work" of the reduction) is some a priori bounded polynomial. A seminal result by Coron (Eurocrypt'02) shows limitations of such reductions; however, his impossibility result and its subsequent extensions all suffer from two notable restrictions: (1) they only rule out so-called "simple" reductions, where the reduction is restricted to only sequentially invoke "straight-line" instances of the adversary; and (2) they only rule out reductions to non-interactive (two-round) assumptions. In this work, we present the first full impossibility result: our main result shows that the existence of any linear-preserving black-box reduction for basing the security of unique signatures on some bounded-round assumption implies that the assumption can be broken in polynomial time.

Hash functions are of fundamental importance in theoretical and in practical cryptography, and with the threat of quantum computers possibly emerging in the future, it is an urgent objective to understand the security of hash functions in the light of potential future quantum attacks. To this end, we reconsider the collapsing property of hash functions, as introduced by Unruh, which replaces the notion of collision resistance when considering quantum attacks. Our contribution is a formalism and a framework that offers significantly simpler proofs for the collapsing property of hash functions. With our framework, we can prove the collapsing property for hash domain extension constructions entirely by means of decomposing the iteration function into suitable elementary composition operations. In particular, given our framework, one can argue purely classically about the quantum-security of hash functions; this is in contrast to previous proofs which are in terms of sophisticated quantum-information-theoretic and quantum-algorithmic reasoning.

Consider the following secret-sharing problem. Your goal is to distribute a long file $s$ between $n$ servers such that $(d-1)$-subsets cannot recover the file, $(d+1)$-subsets can recover the file, and $d$-subsets should be able to recover $s$ if and only if they appear in some predefined list $L$. How small can the information ratio (i.e., the number of bits stored on a server per each bit of the secret) be? We initiate the study of such $d$-uniform access structures, and view them as a useful scaled-down version of general access structures. Our main result shows that, for constant $d$, any $d$-uniform access structure admits a secret sharing scheme with a *constant* asymptotic information ratio of $c_d$ that does not grow with the number of servers $n$. This result is based on a new construction of $d$-party Conditional Disclosure of Secrets (Gertner et al., JCSS '00) for arbitrary predicates over $n$-size domain in which each party communicates at most four bits per secret bit. In both settings, previous results achieved non-constant information ratio which grows asymptotically with $n$ even for the simpler (and widely studied) special case of $d=2$. Moreover, our results provide a unique example for a natural class of access structures $F$ that can be realized with information rate smaller than its bit-representation length $\log |F|$ (i.e., $\Omega( d \log n)$ for $d$-uniform access structures) showing that amortization can beat the representation size barrier. Our main result applies to exponentially long secrets, and so it should be mainly viewed as a barrier against amortizable lower-bound techniques. We also show that in some natural simple cases (e.g., low-degree predicates), amortization kicks in even for quasi-polynomially long secrets. Finally, we prove some limited lower-bounds, point out some limitations of existing lower-bound techniques, and describe some applications to the setting of private simultaneous messages.

The problem of verifiable delegation of computation considers a setting in which a client wishes to outsource an expensive computation to a powerful, but untrusted, server. Since the client does not trust the server, we would like the server to certify the correctness of the result. Delegation has emerged as a central problem in cryptography, with a flurry of recent activity in both theory and practice. In all of these works, the main bottleneck is the overhead incurred by the server, both in time and in space. Assuming (sub-exponential) LWE, we construct a one-round argument-system for proving the correctness of any time $T$ and space $S$ RAM computation, in which both the verifier and prover are highly efficient. The verifier runs in time $n \cdot polylog(T)$ and space $polylog(T)$, where $n$ is the input length. Assuming $S \geq \max(n,polylog(T))$, the prover runs in time $\tilde{O}(T)$ and space $S + o(S)$, and in many natural cases even $S+polylog(T)$. Our solution uses somewhat homomorphic encryption but, surprisingly, only requires homomorphic evaluation of arithmetic circuits having multiplicative depth (which is a main bottleneck in homomorphic encryption) $\log_2\log(T)+O(1)$. Prior works based on standard assumptions had a $T^c$ time prover, where $c \geq 3$ (at the very least). As for the space usage, we are unaware of any work, even based on non-standard assumptions, that has space usage $S+polylog(T)$. Along the way to constructing our delegation scheme, we introduce several technical tools that we believe may be useful for future work.

The Blockwise-Korkine-Zolotarev (BKZ) lattice reduction algorithm is central in cryptanalysis, in particular for lattice-based cryptography. A precise understanding of its practical behavior in terms of run-time and output quality is necessary for parameter selection in cryptographic design. As the provable worst-case bounds poorly reflect the practical behavior, cryptanalysts rely instead on the heuristic BKZ simulator of Chen and Nguyen (Asiacrypt'11). It fits better with practical experiments, but not entirely. In particular, it over-estimates the norm of the first few vectors in the output basis. Put differently, BKZ performs better than its Chen-Nguyen simulation. In this work, we first report experiments providing more insight on this shorter-than-expected phenomenon. We then propose a refined BKZ simulator by taking the distribution of short vectors in random lattices into consideration. We report experiments suggesting that this refined simulator more accurately predicts the concrete behavior of BKZ. Furthermore, we design a new BKZ variant that exploits the shorter-than-expected phenomenon. For the same cost assigned to the underlying SVP-solver, the new BKZ variant produces bases of better quality. We further illustrate its potential impact by testing it on the SVP-120 instance of the Darmstadt lattice challenge.

The RSA PKCS#1 v1.5 signature algorithm is the most widely used digital signature scheme in practice. Its two main strengths are its extreme simplicity, which makes it very easy to implement, and that verification of signatures is significantly faster than for DSA or ECDSA. Despite the huge practical importance of RSA PKCS#1 v1.5 signatures, providing formal evidence for their security based on plausible cryptographic hardness assumptions has turned out to be very difficult. Therefore the most recent version of PKCS#1 (RFC 8017) even recommends a replacement the more complex and less efficient scheme RSA-PSS, as it is provably secure and therefore considered more robust. The main obstacle is that RSA PKCS#1 v1.5 signatures use a deterministic padding scheme, which makes standard proof techniques not applicable. We introduce a new technique that enables the first security proof for RSA-PKCS#1 v1.5 signatures. We prove full existential unforgeability against adaptive chosen-message attacks (EUF-CMA) under the standard RSA assumption. Furthermore, we give a tight proof under the Phi-Hiding assumption. These proofs are in the random oracle model and the parameters deviate slightly from the standard use, because we require a larger output length of the hash function. However, we also show how RSA-PKCS#1 v1.5 signatures can be instantiated in practice such that our security proofs apply. In order to draw a more complete picture of the precise security of RSA PKCS#1 v1.5 signatures, we also give security proofs in the standard model, but with respect to weaker attacker models (key-only attacks) and based on known complexity assumptions. The main conclusion of our work is that from a provable security perspective RSA PKCS#1 v1.5 can be safely used, if the output length of the hash function is chosen appropriately.

Notable recent security incidents have generated intense interest in adversaries which attempt to subvert---perhaps covertly---crypto\-graphic algorithms. In this paper we develop (IND-CPA) Semantically Secure encryption in this challenging setting. This fundamental encryption primitive has been previously studied in the ``kleptographic setting,'' though existing results must relax the model by introducing trusted components or otherwise constraining the subversion power of the adversary: designing a Public Key System that is kletographically semantically secure (with minimal trust) has remained elusive to date. In this work, we finally achieve such systems, even when all relevant cryptographic algorithms are subject to adversarial (kleptographic) subversion. To this end we exploit novel inter-component randomized cryptographic checking techniques (with an offline checking component), combined with common and simple software engineering modular programming techniques (applied to the system's black box specification level). Moreover, our methodology yields a strong generic technique for the preservation of any semantically secure cryptosystem when incorporated into the strong kleptographic adversary setting.

We construct efficient and tightly secure pseudorandom functions (PRFs) with only logarithmic security loss and short secret keys. This yields very simple and efficient variants of well-known constructions, including those of Naor-Reingold (FOCS 1997) and Lewko-Waters (ACM CCS 2009). Most importantly, in combination with the construction of Banerjee, Peikert and Rosen (EUROCRYPT 2012) we obtain the currently most efficient LWE-based PRF from a weak LWE-assumption with a much smaller modulus than the original construction. In comparison to the only previous construction with this property, which is due to Doettling and Schroeder (CRYPTO 2015), we use a modulus of similar size, but only a single instance of the underlying PRF, instead of $\lambda \cdot \omega(\log \lambda)$ parallel instances, where $\lambda$ is the security parameter. Like Doettling and Schroeder, our security proof is only almost back-box, due to the fact that the number of queries made by the adversary and its advantage must be known a-priori. Technically, we introduce all-prefix universal hash functions (APUHFs), which are hash functions that are (almost-)universal, even if any prefix of the output is considered. We give simple and very efficient constructions of APUHFs, and show how they can be combined with the augmented cascade of Boneh et al. (ACM CCS 2010) to obtain our results. Along the way, we develop a new and more direct way to prove security of PRFs based on the augmented cascade.

Software side channel attacks have become a serious concern with the recent rash of attacks on speculative processor architectures. Most attacks that have been demonstrated exploit the cache tag state as their exfiltration channel. While many existing defense mechanisms that can be implemented solely in software have been proposed, these mechanisms appear to patch specific attacks, and can be circumvented. In this paper, we propose minimal modifications to hardware to defend against a broad class of attacks, including those based on speculation, with the goal of eliminating the entire attack surface associated with the cache state covert channel. We propose DAWG, Dynamically Allocated Way Guard, a generic mechanism for secure way partitioning of set associative structures including memory caches. DAWG endows a set associative structure with a notion of protection domains to provide strong isolation. When applied to a cache, unlike existing quality of service mechanisms such as Intel's Cache Allocation Technology (CAT), DAWG isolates hits and metadata updates across protection domains. We describe how DAWG can be implemented on a processor with minimal modifications to modern operating systems. We argue a non-interference property that is orthogonal to speculative execution and therefore argue that existing attacks such as Spectre Variant 1 and 2 will not work on a system equipped with DAWG. Finally, we evaluate the performance impact of DAWG on the cache subsystem.