Post-quantum cryptographic schemes like ML-KEM and ML-DSA have been standardized to secure digital communication against quantum threats. While their theoretical foundations are robust, we identify a critical implementation-level vulnerability in both: a single point of failure centered on the random seed pointer used in polynomial sampling. By corrupting this pointer, an attacker can deterministically compromise the entire scheme, bypassing standard countermeasures. We present the first practical fault-injection attacks exploiting this weakness and validate them on an STM32H7 microcontroller using laser fault injection. Our results demonstrate full key and message recovery for ML-KEM and signature forgery for ML-DSA, with success rates up to 100%. We further verify the presence of this vulnerable implementation style in widely used public libraries, including PQM4, LibOQS, PQClean, and WolfSSL, and propose effective countermeasures to mitigate this overlooked yet severe threat.

We build two-server private information retrieval (PIR) that achieves information-theoretic security and strong double-efficiency guarantees. On a database of $n > 10^6$ bits, the servers store a preprocessed data structure of size $1.5 \sqrt{\log_2 n} \cdot n$ bits and then answer each PIR query by probing $12 \cdot n^{0.82}$ bits in this data structure. To our knowledge, this is the first information-theoretic PIR with any constant number of servers that has quasilinear server storage $n^{1+o(1)}$ and polynomially sublinear server time $n^{1-\Omega(1)}$. Our work builds on the PIR-with-preprocessing protocol of Beimel, Ishai, and Malkin (CRYPTO 2000). The insight driving our improvement is a compact data structure for evaluating a multivariate polynomial and its derivatives. Our data structure and PIR protocol leverage the fact that Hasse derivatives can be efficiently computed on-the-fly by taking finite differences between the polynomial's evaluations. We further extend our techniques to improve the state-of-the-art in PIR with three or more servers, building on recent work by Ghoshal, Li, Ma, Dai, and Shi (TCC 2025). On an 11 GB database with 1-byte records, our two-server PIR encodes the database into a 1 TB data structure – which is 4,500,000$\times$ smaller than that of prior two-server PIR-with-preprocessing schemes, while maintaining the same communication and time per query. To answer a PIR query, the servers fetch and send back 4.4 MB from this data structure, requiring 2,560$\times$ fewer memory accesses than linear-time PIR. The main limitation of our protocol is its large communication complexity, which we show how to shrink to $n^{0.31} \cdot \mathsf{poly}(\lambda)$ using compact linearly homomorphic encryption.

We review $k$-anonymity in authentication schemes, group signatures, and ring signatures. Existing constructions either require a signer to maintain state across interactions, or admit tracing algorithms that cost $O(n^k)$ in the number $n$ of signatures. We introduce $k$-anonymous group signatures ($k$-AGS), that achieves stateless signing with a tracing cost $O(n+k)$, while necessarily sacrificing unlinkability. We present a generic construction of $k$-AGS together with an efficient instantiation. We additionally construct a proof-of-concept $k$-unlinkable group signature ($k$-UGS) scheme that achieve unlinkability at the expense of $O(n^2k)$ tracing overhead. We pose the question of whether stateless signing, $O(n+k)$ tracing, and unlinkability are simultaneously achievable, and leave it as an open problem. Building on our $k$-AGS framework, we introduce $k$-Anonymous Set Pre Constrained Group Signatures ($k$-ASPCGS), a threshold variant of Set Pre-Constrained Group Signatures introduced by Bartusek et al. (EUROCRYPT '23). As a proof-of-concept, we present two generic constructions of $k$-ASPCGS. We show that our notions arise naturally in the context of content moderation in end-to-end encrypted messaging platforms, where users must remain anonymous until a threshold number of distinct illegitimate contents have been reported.

We present a simple and efficient Byzantine Agreement protocol in the mixed fault model where up to $t$ parties can be Byzantine, up to $s$ parties can be send-omission, and up to $r$ parties can be receive-omission such that $2t+s+r<n$. Our synchronous protocol greatly improves over the efficiency of the state-of-the-art solution due to Loss and Stern [TCC '23]. Specifically, our protocol incurs an expected communication complexity of $O(n^2)$ instead of $O(n^5)$ in their construction, while maintaining the same resilience. Our protocol terminates in an expected constant number of rounds, provided that a constant fraction of parties are non-faulty.

We examine the relationship between correctness definitions for Fully Homomorphic Encryption (FHE) and the associated security definitions. We show that reactive notions of correctness imply INDCPA-D and sINDCPA-D security. But that to obtain both INDCPA-D and sINDCPA-D security we need to use a randomized version of the evaluation procedure. Such randomized evaluation procedures cause problems in real life deployments of FHE solutions, so we then go on to show how one can de-randomize the evaluation procedure and still obtain sINDCPA-D security in the random oracle model for the specific FHE scheme of TFHE.

Guo, Li, and Qin proposed a lightweight certificateless encryption (CLE) scheme designed for IoT environments (\textit{Discover Computing}, 2025). This paper demonstrates that the proposed scheme does not achieve CCA security, contrary to the authors' claim. Specifically, we identify two critical points. First, since the ciphertext retains a multiplicative ElGamal structure, it can always be re-randomized using arbitrary randomness. Second, based on this property, an adversary can transform a challenge ciphertext into another valid ciphertext of the same plaintext, and then query the decryption oracle with the transformed ciphertext to recover the challenge plaintext. This attack exploits a definitional gap in the CCA game, where only direct decryption queries on the challenge ciphertext are prohibited. In this work, we formalize the attack procedure and verify its validity based on implementation.

The Hamming Quasi-Cyclic (HQC) scheme has recently been standardized as a post-quantum key encapsulation mechanism (KEM), emphasizing the importance of efficient and secure hardware realizations on embedded platforms. However, HQC relies heavily on sparse–dense polynomial multiplications, where conventional shift-and-add architectures remain both performance- and security-critical. In FPGA implementations, these multiplications dominate execution time—occupying 59.5%, 56.1%, and 58.3% of the total latency for KeyGen, Encap, and Decap, respectively—and are further vulnerable to correlation power analysis (CPA) due to deterministic, index-driven memory access patterns. As countermeasures, parallelization improves performance at the cost of additional area. Dummy insertion with random shuffling mitigates leakage but incurs extra cycle overhead. To address this, we propose a co-designed dummy-inserted parallel shift-and-add multiplier for HQC. The design integrates dummy insertion and two-index parallelism in a complementary manner, achieving reduced cycles with area efficiency while providing intrinsic resistance to CPA. Implemented on a Xilinx Artix-7 FPGA, the proposed architecture achieves up to a 1.25× speedup over the baseline sequential multiplier while maintaining near–state-of-the-art area–time efficiency—incurring only a 1.16× AT overhead to simultaneously deliver accelerated performance and CPA resistance. Test Vector Leakage Assessment (TVLA) measurements and theoretical analysis confirm that the parallel architecture effectively suppresses power-based side-channel leakage and provides inherent resistance against CPA—reducing significant leakage points from 4.29% to 0.09%. This work demonstrates that performance and side-channel resistance can be jointly optimized through synergistic hardware–algorithm co-design, offering a practical and scalable HQC accelerator for post-quantum embedded systems.

We construct a pseudorandom correlation function (PCF) for oblivious linear evaluation (OLE) from sparse LPN over any finite field. The programmability property of our PCF implies a PCF for any multiparty degree-two correlation, e.g., Beaver triples. Our PCF is the first PCF for degree-two correlations from a well-established cryptographic assumption, apart from (inefficient) generic PCFs based on homomorphic secret sharing or fully homomorphic encryption. Our PCF outperforms the previously fastest PCF for Beaver triples (Boyle et al., Crypto 2022) by 3.2-28x. We build on the recent pseudorandom correlation generator (PCG) by Miao et al. (Asiacrypt 2025) and extend it to a PCF using a recursive approach similar to Braun et al. (Asiacrypt 2025). Moreover, we extend these techniques to support authenticated degree-two correlations in the important two-party case.

Verifiable Secret Sharing (VSS) schemes are fundamental building blocks in distributed cryptography. While most existing works focus on threshold structures, many real-world applications require more general access structures, where participants have different levels of power and only certain subsets are authorized to reconstruct the secret. Existing computational VSS schemes for general access structures typically rely on Discrete Logarithm (DL)-based homomorphic commitments, which limits their applicability, particularly in scenarios requiring Post-Quantum (PQ) security. In this work, we present a generalized version of $\mathrm{\Pi}$, a unified framework introduced at PKC 2025 for constructing computational VSS schemes without relying on homomorphic commitments. Our framework supports arbitrary monotone $\mathcal{Q}_2$ access structures, encompassing replicated and threshold secret sharing (e.g., Shamir's scheme), while preserving the efficiency and modularity of $\mathrm{\Pi}$. Notably, it requires only a random oracle and any commitment scheme satisfying hiding and binding, making it compatible with a wide range of instantiations, including PQ-secure commitments. In particular, our hash-based instantiation yields the first symmetric-key-based VSS scheme for general access structures. Compared to prior general-access VSS schemes based on homomorphic commitments (e.g., variants of Pedersen scheme from FC 2003), our DL-based constructions eliminate the need for homomorphic commitments and achieve asymptotic improvements in verification and reconstruction costs. We believe that this extension enhances the versatility of the original $\mathrm{\Pi}$ framework and paves the way for its deployment in a broader range of practical distributed systems.

To mitigate trust concerns in the setup phase of pairing-based zk-SNARKs, the primary solution has been the sampling of the Structured Reference String (SRS) using an MPC protocol. In 2017, Bowe, Gabizon, and Miers introduced the Powers of Tau MPC protocol for sampling a universal SRS, which has since become the main SRS generation protocol for numerous practical projects. The protocol's designers showed that for a circuit with $2^{21}$ multiplication gates, verifying the universal SRS for Groth16 zk-SNARK could take $55$ minutes for a single update. However, they clarified that "the verification is not run by individual users; it is done by the coordinator and anyone who wishes to verify the transcript of the protocol after completion". This note demonstrates the importance of verifying the final SRS by either $\textit{each}$ individual end-user or $\textit{all}$ ceremony participants to mitigate potential attacks. We discuss simple attack scenarios that highlight vulnerabilities if $\textit{each}$ end-user or $\textit{all}$ participants fail to verify the final SRS. Additionally, by leveraging batching and aggregating techniques, we introduce an efficient verification algorithm for the (original) Powers of Tau protocol, substantially reducing SRS verification time and making it practical even for large-scale ceremonies. In the case of rejection, a more efficient recursive verification approach aids in identifying malicious parties more effectively. This note aims to enhance procedural understanding of SRS generation ceremonies through the Powers of Tau protocol and improve the reliability of current ceremonies against potential threats.

The MPC-in-the-Head paradigm is a promising approach for constructing post-quantum signature schemes. Its significance is underscored by NIST's selection of six signatures based on this paradigm and its variants, TC-in-the-Head and VOLE-in-the-Head, among the fourteen round-2 candidates in its additional post-quantum cryptography standardization process. Recent works by Aguilar-Melchor et al. (ASIACRYPT 2023), Hülsing et al. (CRYPTO 2024), and Baum et al. (CRYPTO 2025) have established EUF-CMA security for these signatures in the Quantum Random Oracle Model (QROM). However, their proofs do not account for crucial optimization techniques such as rejection sampling and grinding, rendering them inapplicable to practical schemes like the NIST round-2 candidates Mirath and RYDE. This paper addresses this gap by analyzing the QROM security of MPC-in-the-Head signatures that incorporate these optimizations, with a focus on Mirath and RYDE. We make two main contributions: 1) We provide a new (strong) EUF-CMA security proof that accommodates rejection sampling and grinding. We also present a new EUF-NMA security proof compatible with these optimizations, by extending the techniques of Don et al. (CRYPTO 2022) and Aguilar-Melchor et al. (ASIACRYPT 2023). 2) We also point out a gap in the EUF-CMA security proof of the MPC-in-the-Head signature schemes using correlated-tree techniques, MQOM, SBC (Huth and Joux, CRYPTO 2024), and rBN++ (Kim, Lee, and Son, EUROCRYPT 2025).

As three frequently used techniques for adaptive reprogramming in the QROM, the adaptive One-Way to Hiding (O2H) proposed by Unruh (CRYPTO 2014), the GHHM adaptive reprogramming proposed by Grilo et al. (ASIACRYPT 2021), and the Pan-Zeng adaptive reprogramming proposed by Pan and Zeng (PKC 2024), address different reprogramming scenarios, and do not appear to imply one another. A recent breakthrough by Jaeger (ASIACRYPT 2025) reveals a surprising connection: all three of these adaptive techniques can be implied by a non-adaptive reprogramming technique called Fixed-Permutation O2H (FP-O2H). Furthermore, Jaeger's result also improves the security bounds for Unruh's adaptive O2H and the Pan-Zeng adaptive reprogramming theorem. In this paper, we reconsider the implication between FP-O2H and GHHM adaptive reprogramming. We first introduce a variant of FP-O2H, called the Double-Oracle-Fixed-Permutation O2H (DOFP-O2H). Then, by applying this variant, we derive a tighter upper bound for the GHHM adaptive reprogramming. Thereby, our result complements Jaeger’s findings by addressing the final piece, showing that the non-adaptive O2H not only implies adaptive reprogramming in the QROM but also yields tighter upper bounds. In addition, a direct application of our tighter GHHM adaptive reprogramming yields a tighter \textsf{EUF-CMA} security proof of the Fiat–Shamir transform in the QROM: the security loss with respect to the number of signing queries q_s decreases from O(q_s) to O(\sqrt{q_s}). Furthermore, we reconsider the implication between FP-O2H and the ABKM permutation resampling proposed by Alagic et al. (EUROCRYPT 2022). By applying our DOFP-O2H, we reprove the ABKM permutation resampling theorem, and derive the same upper bound as that of Alagic et al. This result suggests that the FP-O2H not only can be applied to analyze the reprogramming in the QROM, but also has potential for analyzing reprogramming in the random permutation setting.

Recently, Hövelmanns, Hülsing, and Majenz introduced a security notion called Find Failing Plaintext – Non Generic (FFP-NG), which captures the ability of an adversary to find decryption failures by making non-trivial use of the public key. A first analysis of this property for lattice-based schemes was presented by Majenz and Sisinni, who showed that the Learning With Errors (LWE) problem reduces to breaking the FFP-NG security of the PVW scheme with discrete Gaussian noise. In this work, we generalize their result by analysing the FFP-NG security of widely used schemes based on Ring-LWE and Module-LWE. To keep our analysis as general as possible, we consider a family of subgaussian distributions that includes, among others, discrete Gaussians and centered binomials.

Key-dependent attacks are effective only for specific weak-key classes, limiting their practical impact. We present a generic statistical framework that combines multiple key-dependent distinguishers into universal attacks covering the full key space. Using log-likelihood ratio statistics, our framework tests the secret key against multiple weak-key distinguishers, aggregates their evidence to determine whether the key is weak or strong for each distinguisher, and exploits this classification to reduce the effective key entropy for key recovery. We apply this to Orthros-PRF, a sum-of-permutations (SoP) design where any differential-based distinguisher holds only for a fraction of keys. This yields the first universal 8-round differential-linear (DL) key-recovery attack with median time complexity $2^{119.58}$, whereas prior work reached at most 7 rounds in the weak-key setting. To discover the required distinguishers, we extend the open-source S-box Analyzer tool with MILP support for deterministic propagation and develop a model integrating distinguisher search with key recovery. This enables automated discovery of multidimensional DL distinguishers covering up to 10 rounds in each Orthros branch, improving prior work by 4 rounds. Our results demonstrate that statistical aggregation of multiple weak-key distinguishers enables effective universal cryptanalysis. Our framework is generic and is applicable to other primitives with multiple identifiable weak-key classes.

Anonymous credentials enable unlinkable and privacy-preserving user authentication. To ensure non-transferability of credentials among corrupt users, they can additionally be device-bound. Therein, a credential is tied to a key protected by a secure element (SE), usually a hardware component, and any presentation of the credential requires a fresh contribution of the SE. Interestingly, despite being a fundamental aspect of user credentials, device binding for anonymous credentials is relatively unstudied. Existing constructions either require multiple calls to the SE, or need the SE to keep a credential-specific state -- violating core design principles of shielded SEs. Further, constructions that are compatible with the most mature credential scheme BBS rely on the honesty of the SE for privacy, which is hard to vet given that SEs are black-box components. In this work, we thoroughly study and solve the problem of device-bound anonymous credentials (DBACs). We model DBACs to ensure the unforgeability and non-transferability of credentials, and to guarantee user privacy at the same time. Our definitions cover a range of SE trust levels, including the case of a subverted or fully corrupted SE. We also define blind DBACs, in which the SE learns nothing about the credential presentations it helped compute. This targets the design of a remote, cloud-based SE which is a deployment model considered for the EU Digital Identity (EUDI) wallet to address the fact that most user phones are not equipped with a sufficiently secure SE. Finally, we present three simple and round-optimal constructions for device binding of BBS credentials, and prove their security in the AGM+ROM and privacy unconditionally. The SE therein is extremely lightweight: it only has to compute a BLS or Schnorr signature in a single call. We also give the BLS-based construction in a blind variant, yielding the first protocol that enables privacy-preserving device binding for anonymous credentials when being used with a remote SE.

Pseudorandom generators (PRGs) are a foundational primitive in classical cryptography, underpinning a wide range of constructions. In the quantum setting, pseudorandom quantum states (PRSs) were proposed as a potentially weaker assumption that might serve as a substitute for PRGs in cryptographic applications. Two primary size regimes of PRSs have been studied: logarithmic-size and linear-size. Interestingly, logarithmic PRSs have led to powerful cryptographic applications, such as digital signatures and quantum public-key encryption with tamper-resilient keys, that have not been realized from their linear counterparts. However, PRGs have only been black-box separated from linear PRSs, leaving open the fundamental question of whether PRGs are also separated from logarithmic PRSs. In this work, we resolve this open problem. We establish a quantum black-box separation between (quantum-evaluable) PRGs and PRSs of either size regime. Specifically, we construct a unitary quantum oracle with inverse access relative to which no black-box construction of PRG from (logarithmic or linear) PRS exists. This does not directly separate PRG from some of the applications of SPRS since these applications involve, as a first step, a non-black-box construction of a notion termed bot-PRGs. To address this, we present another unitary separation showing that PRG are also separated from bot-PRGs. Thus, we obtain separation from digital signatures and quantum public-key encryption.

The FRI protocol (ICALP '18) is one of the most influential and widely deployed building blocks at the core of modern SNARK systems. While its concrete security is well understood, existing security proofs are intricate and technically complex. In this work, we present a significantly simpler security analysis of FRI, in particular its round-by-round soundness. Our approach is more accessible to a broader audience, lowering the barrier to understanding this fundamental protocol. Furthermore, the simplicity of our analysis may pave the way for future formal verification efforts of modern SNARK constructions.

Since the work of Chaum in ’82, the problem of designing secure blind signature protocols for existing signature schemes has been of great interest. In particular, when considering Schnorr signatures, nowadays used in Bitcoin, designing corresponding efficient and secure blind signature schemes is very challenging in light of the ROS attack (Eurocrypt’21), which affected all previous efficient constructions. Currently, the main positive result about concurrent-secure blind Schnorr signatures is the one of Fuchsbauer and Wolf (Eurocrypt’24). Their construction is quite demanding, indeed it requires trusted parameters, non-interactive zero-knowledge arguments and CPA-secure public-key encryption. Moreover, it is proven secure under a game-based definition only, is limited to computational blindness and is vulnerable to harvest now “link” later quantum attacks. Nicely, their construction is also a predicate blind signature (PBS) scheme, allowing signers to have some partial control on the content of the blindly signed message. In this work, we show neat improvements to the state-of-the-art presenting a new construction for concurrent-secure blind Schnorr signatures that relies on milder/reduced cryptographic assumptions, enjoys statistical blindness, replaces the problematic trusted setup with a non-programmable random oracle (NPRO), and satisfies also a one-sided simulation-based property providing deniability in PBSs.

The post-quantum signature scheme Falcon is an attractive scheme for constrained devices due to its compactness and verification performance. However, it relies on floating-point arithmetic for signature generation, which - alongside physical security concerns - introduces two additional drawbacks: Firstly, if implemented using the standard double-precision format, Falcon does not satisfy the formally proven error bounds required for a secure Gaussian sampler implementation. Although no practical attacks exploiting this limitation are currently known, it does give future attack concerns. Secondly, when looking at constrained devices, 32-bit constrained devices can lack hardware support for high-precision floating-point arithmetic and its use introduces significant performance overhead, as it must be emulated using integers. In this work we present a novel method to address these limitations: We show that Falcon can be implemented using $\textit{single-precision}$ floating-point numbers. Our proposed method uses Triple-Word Floating-Point (TW) arithmetic and achieves a precision of at least 72 bits, compared to the 53 bits of double-precision floating-point arithmetic. We show our implementation achieves error bounds that meet the formal security requirements for a secure Gaussian sampler implementation, while maintaining other security guarantees. This way, Falcon can run on constrained devices equipped only with a single-precision Floating-Point Unit (FPU) without the need for integer emulation. We demonstrate the feasibility of our approach on the Nucleo-L4R5ZI board, which features a Cortex-M4F processor enabled with a single-precision FPU. More precisely, we show the cost of increasing the precision of Falcon in this way only increases the computational effort by a factor of approximately 1.84 compared to the CPU cycles required for an implementation using emulated double-precision arithmetic via integers.

Although the lattice-estimator predicts that Learning with Errors instances having small and very sparse secrets can be broken by hybrid attacks with modest computational resources, no efficient open-source implementation of these attacks exists. This work implements the so-called Guess + Verify attack (G+V) analysed by Albrecht et al. (SAC'19), containing three improvements: (1) cuBLASter, a GPU-based implementation of the lattice basis reduction software BLASter by Ducas et al. (ASIACRYPT'25); (2) a dimension reduction technique for the BDD instance; and (3) a batched variant of Babai’s Nearest Plane algorithm. On bases of dimension 512 and above, cuBLASter outperforms BLASter. We also integrate the GPU implementation of the General Sieve Kernel by Ducas et al. (EUROCRYPT'21) into cuBLASter’s BKZ framework. Running G+V on the benchmark instances by Wenger et al. (IEEE SP'25), we show that G+V achieves significantly higher success rates than the Cool&Cruel attack (C+C) by Nolte et al. (AFRICACRYPT'24) on almost all instances, and G+V's average CPU and GPU utilization is substantially lower than the minimum reported by C+C.

Indistinguishability Obfuscation (iO) renders indistinguishable any pair of same-sized and functionally identical circuits. Since its introduction by Barak et al. (Crypto ’01), iO has become the holy grail of modern cryptography due to the myriad of applications it enables. Despite several theoretical constructions (including recent breakthrough results from well-studied assumptions), practically efficient frameworks for iO have remained out of reach. In this paper, we formally introduce the notion of Hardware-based indistinguishability obfuscation (HiO) with an aim to make iO practical. Further, we present a novel framework called Hardware-based Circuit Obfuscation from Data Encryption (HardCODE) for realizing highly efficient HiO. Given a simple tamper-proof hardware to store a fixed size secret key, our proposed HardCODE framework yields a provably secure hardware module that achieves iO for all poly-sized circuits, while additionally relying on any one-way function and certain (heuristic but widely studied) assumptions on substitution-permutation networks (used extensively in the design of secure block ciphers). We practically instantiate HardCODE from widely deployed realizations of a simple HSM (Hardware Security Module), and the substitution-permutation network underlying the PRESENT block cipher (an ISO standard for lightweight cryptography). We then use this instance of HardCODE to obfuscate, without loss of generality, the C-17 circuit of the ISCAS-85 benchmark suite, while incurring very small practical overheads. To the best of our knowledge, this constitutes the first practical demonstration of iO.

NTRU is one of the most extensively studied lattice-based cryptographic schemes and is widely regarded as a strong candidate for post-quantum security. The most effective attacks on NTRU are lattice-based or lattice-related, which naturally guide the choice of parameters to achieve the desired security levels. In 1997, Hoffstein and Silverman proposed a variant of NTRU based on a noncommutative algebraic structure, claiming that it would mitigate lattice attacks. However, their scheme was later shown to be vulnerable to an algebraic attack by Coppersmith. Although several subsequent attempts have been made in the literature to develop noncommutative variants of NTRU, most of these designs have either been shown to be vulnerable to algebraic attacks or have failed to directly address lattice-based attacks. In this work, we revisit the problem of constructing a noncommutative analog of NTRU that offers stronger resistance against direct lattice attacks. Firstly, we conceptualize the problem by introducing an almost unstructured variant, and then refine this idea towards a more compact instantiation by introducing a semi-structured variant. Our proposal may be viewed as a follow-up to the early noncommutative construction of Hoffstein and Silverman. We further discuss the requirements towards having a fully structured construction in this paradigm.

As the Hamming Quasi-Cyclic (HQC) cryptosystem was recently selected by NIST for standardization, a thorough evaluation of its implementation security is critical before its widespread deployment. This paper presents single-trace side-channel attacks that recover the full long-term secret key of HQC, experimentally evaluated on a protected Cortex-M4 implementation. We introduce two distinct attacks that significantly advance the state of the art: a passive attack that uniquely models key recovery as a moderate-density parity-check (MDPC) decoding problem from a single valid ciphertext, and an active chosen-ciphertext attack employing a new probing strategy on a linear combination of secret key components for significantly improved efficiency. Both attacks are enabled by a new information set decoding (ISD) variant that exploits soft side-channel information, a contribution of broader importance to code-based cryptography. Our experiments show that a single trace suffices for full key recovery under realistic conditions, effectively defeating countermeasures such as codeword masking for the first time. We also show that several existing defenses are ineffective against the new attacks.

In NDSS 2024, Yu~et al. proposed AAKA, an Anonymous Authentication and Key Agreement scheme designed to protect users' privacy from mobile tracking by Mobile Network Operators (MNOs). AAKA aims to provide both anti-tracking privacy and traceability (lawful de-anonymization), allowing subscribers to access the network via anonymous proofs while enabling a Law Enforcement Agency (LEA) to trace the real identity if misbehaviors are detected. However, we identify that the AAKA scheme in NDSS 2024 is insecure since the subscriber's identity is exposed within the protocol, thereby failing to achieve the claimed privacy and traceability. Building on the repair of AAKA, we propose AAKA+, Anonymous Authentication and Key Agreement with Verifier-Local Revocation, a new mobile authentication scheme, to ensure privacy against mobile tracking. In addition to the privacy and traceability introduced in NDSS 2024, AAKA+ additionally allows the MNO to immediately assert whether the associated subscriber has been traced and revoked upon receiving an anonymous proof. We formally define the syntax and the security model of AAKA+ and propose two concrete schemes, AAKA+BB and AAKA+PS, based on the Boneh-Boyen signature and the Pointcheval-Sanders signature schemes, respectively. Both AAKA+BB and AAKA+PS are pairing-free on the user equipment side and compatible with existing cellular infrastructure. Experimental results show that our schemes are practical, with anonymous proof generation taking approximately 18 milliseconds for a constrained device.

The growing need for secure computation has spurred interest in cryptographic techniques that operate on encrypted data without revealing its content. Fully Homomorphic Encryption (FHE), especially LWE-based schemes, enables such processing while preserving confidentiality. Decentralized computing offers scalable resources without requiring in-house servers, but it relies heavily on the confidentiality guarantees of underlying schemes. While many existing protocols successfully protect input privacy, function confidentiality remains a largely overlooked but crucial aspect of secure delegated computation. In this work, we present a novel Oblivious Universal Function (OUF) scheme that enables a client to outsource computation of an arbitrary function to an untrusted server while hiding both the input data and the function being applied. Our construction leverages LWE-based FHE and a virtual-tape evaluation model to support composable, non-interactive, and reusable function execution. Crucially, the server remains oblivious not only to the encrypted inputs but also to the structure, type, and identity of the function it is executing. OUF thus bridges the gap between theoretical privacy guarantees and practical secure computation in decentralized environments.

In a secret sharing scheme for a monotone access structure $\mathcal{A}:\{0,1\}^n\rightarrow \{0,1\}$, a dealer can share a secret $s$ to $n$ parties such that any authorized subset of parties $A\in\mathcal{A}$ can recover $s$ while all other subsets learn nothing about $s$. In this work, we study fully anonymous secret sharing (FASS), which strengthens standard secret sharing by requiring the following properties: - Share Anonymity. The shares belonging to any unauthorized set of parties not only hide the secret, but also all identifiable information such as party identities and whether or not the shares were generated together. In particular, it suffices that such shares be uniform and independent. - Anonymous Reconstruction. The reconstruction algorithm does not need to know the reconstructing set of parties. Efficient FASS exists for threshold access structures. For general access structures, the only known construction relies on a monotone DNF representation of $\mathcal{A}$ and has per-party share size $\Omega(\ell n)$ where $\ell$ is the number of minterms of $\mathcal{A}$. This leaves an exponential gap between standard secret sharing and FASS even for simple access structures. Moreover, even in the threshold case, known schemes could not achieve optimal robust reconstruction when mixing shares of multiple secrets. Motivated by a recent work of Eldridge et al. [USENIX'24], who demonstrated an application of FASS to stalker detection, we initiate a systematic study of FASS, obtaining the following main results. - Near-Optimal Information-Theoretic FASS. We obtain strong lower bounds, showing that the dependence on the DNF size is generally inherent. In particular, the share size can be exponential in the number of parties or even in the minimum CNF size. This stands in sharp contrast to standard secret sharing, where no super-polynomial lower bounds are known, and where the share size is upper bounded by the CNF size. For DNF with $\ell$ small minterms, we improve the previous upper bound to $\tilde O(\ell)$, matching our lower bound up to a polylogarithmic factor. - Computational FASS. We show that the above negative results can be circumvented in the computational setting, obtaining FASS schemes with succinct shares. Under the learning with errors (LWE) assumption, we present a general compiler from standard secret sharing to FASS that preserves the share size of the underlying scheme. For natural graph access structures, we directly construct succinct FASS from either one-way functions or bilinear maps. - Robust FASS. We show that simple modifications of our computational FASS schemes can allow for robust reconstruction of a polynomially unbounded number of secrets from any mixture of their authorized shares.

We present SCORE, a tweaked version of the SlotToCoeff operation tailored to the context of encrypted real vectors in the CKKS scheme, where SCORE stands for “SlotToCoeff Optimization for Real-Vector Encryption”. This approach accelerates CKKS bootstrapping algorithms that employ the SlotToCoeff operation as their final step, provided the inputs are encryptions of real vectors. SCORE constitutes a conceptually simple refinement of the conventional SlotToCoeff step that can be seamlessly incorporated into existing bootstrapping algorithms. Despite its minimal nature, the technique yields tangible performance improvements. We demonstrate its utility through proof-of-concept implementations for two such algorithms: the conventional bootstrapping algorithm and the SPRU algorithm, which is particularly efficient for ciphertexts containing a small number of encrypted slots. Our results indicate notable performance gains of up to $2\times$ compared to the original formulations of the algorithms.

Anonymous communication networks (ACNs) aim to thwart an adversary, who controls or observes chunks of the communication network, from determining the respective identities of two communicating parties. We focus on low-latency ACNs such as Tor, which target a practical level of anonymity without incurring an unacceptable transmission delay. While several definitions have been proposed to quantify the level of anonymity provided by high-latency, message-centric ACNs (such as mix-nets and DC-nets), this approach is less relevant to Tor, where user–destination pairs communicate over secure overlay circuits. Moreover, existing evaluation methods of traffic analysis attacks on Tor appear somewhat ad hoc and fragmented. We propose a fair evaluation framework for such attacks against onion routing systems by identifying and discussing the crucial components for evaluation, including how to consider various adversarial goals, how to factor in the adversarial ability to collect information relevant to the attack, and how these components combine to suitable metrics to quantify the adversary's success.

Anonymous credentials enable the unlinkable presentation of previously attested information, or even only predicates thereof. They are a versatile tool and currently enjoy attention in various real-world applications, ranging from the European Digital Identity project to Privacy Pass. While each application usually requires their own tailored variant of anonymous credentials, they all share the same common blueprint. So far, this has not been leveraged though, and currently several proposals either targeting monolithic variants of core components such as BBS signatures, or application-specific protocols undergo standardization. This is clearly not optimal, as the same work gets repeated multiple times, while still risking ending up with many slight modifications of the same main idea and protocols. In this work we present our vision to use a modular approach to build anonymous credential systems: they are built from a core component – consisting of a commitment, signature and NIZK scheme – that can be extended with additional commitment-based modules in a plug-and-play manner. We sketch modules for pseudonyms, range proofs and device binding. Importantly, apart from the committed input, all modules are entirely independent of each other. We use this modularity to propose a concrete instantiation that uses BBS signatures for the core component and ECDSA signatures for device binding, addressing the need to bind modern credential schemes to legacy signatures in secure hardware elements.

In a secret sharing scheme for a monotone access structure $\mathcal{A}$, one can share a secret among a set of parties such that all subsets of parties authorized by $\mathcal{A}$ can reconstruct the secret while all other subsets learn nothing. However, what if an unauthorized subset of parties collude and offer their shares for sale? Specifically, suppose that the parties pool their shares to create a reconstruction box that reveals the secret upon receiving enough additional shares as input. To deter this behavior, Goyal et al. (CRYPTO'21) introduced the notion of traceable secret sharing (TSS), where it is possible to provably trace reconstruction boxes containing leaked secret shares back to their respective parties. Goyal et al. and subsequent work presented definitions and constructions of TSS for the threshold access structure. In this work, we revisit the notion of TSS. - We identify shortcomings in previous formulations of TSS and present new, strengthened definitions that are not achieved by known constructions. We show that it is easy to build reconstruction boxes for which the tracing guarantees of all previous works fail. - We extend the study of TSS beyond threshold to general access structures. Our new definitions are, in fact, necessary for this setting as natural adaptations of existing definitions become vacuous for general access structures. - We present new constructions of TSS that satisfy our definitions for all access structures captured by monotone circuits. One of our constructions relies solely on one-way functions while the other additionally relies on indistinguishability obfuscation.

Singh et. al. recently uploaded a preprint describing a new hash function inspired by the Collatz Conjecture. After performing some cursory tests, the proposed function appears to be completely unsuitable for cryptographic purposes, and should not be used.

We initiate a systematic study of information-theoretic cryptography with {\em weak privacy}, only requiring that the adversary cannot rule out any possible secret. For a parameter $0<p\le 1$, we say that a primitive has $p$-weak privacy ($p$-WP) if for every possible view $V$ and inputs $x_1,x_2$, the ratio between the probabilities of the adversary observing $V$ on $x_1$ and on $x_2$ is at least $p$. This generalizes both perfect privacy, which is $p$-WP for $p=1$, and a previous ``combinatorial'' notion of privacy, which is $p$-WP for {\em some} $p>0$. We obtain the following main results. Positive results. We present efficient WP constructions for generalized secret sharing, decomposable randomized encodings, and the related notions of garbling schemes and PSM protocols, as well as interactive secure multiparty computation protocols in the plain model and in the OT-hybrid model. For secret sharing, we settle a question of Beimel and Franklin (TCC 2007), showing that every $n$-party access structure admits a WP scheme with per-party share size $O(n)$. When all unauthorized sets have constant size, we get a $p$-WP scheme with constant share size and $p\ge 1/poly(n)$. Negative result. For decomposable randomized encodings, we show that a previous lower bound technique of Ball et al.\ (ITCS 2020) applies also to the WP notion. Together with our upper bound, this shows that the optimal WP garbling size of the worst-case $f:\{0,1\}^n\to\{0,1\}$ is $\tilde{\Theta}(n^2)$. Application. While WP may seem like an unrealistically weak security notion, we demonstrate its usefulness towards achieving traditional security guarantees. Concretely, under the standard LPN assumption, we show that any $p$-WP secret-sharing scheme with inverse-polynomial $p$ implies a {\em computationally secure} secret-sharing scheme for a related access structure. Together with our positive results for WP secret sharing, this implies a super-polynomial improvement of the share size for a natural class of access structures.

The ARADI block cipher is developed by the U.S. National Security Agency (NSA) as part of its efforts to design secure and efficient cryptographic algorithms. In this paper, we present the first detailed analysis of the lightweight block cipher ARADI under differential fault attacks. Although ARADI is structured around word-wise operations, its security fundamentally depends on bit-level properties, making it vulnerable to carefully crafted fault models. Bit-based fault models require a large number of faults, while word-based models reduce the number of faults but make precise identification of fault locations difficult. To overcome these limitations, we introduce a new nibble-based approach using permissible nibble differences (PNDs). This model provides an effective tradeoff between the number of required faults and the complexity of identifying fault positions. In our attack, the adversary randomly injects nibble faults before the last two rounds and constructs differential equations involving unknown key variables without knowing the exact fault values. Our results demonstrate that, with at most 108 random fault injections, the entire master key of ARADI can be recovered within a practical time complexity. This work reveals previously unexplored vulnerabilities of ARADI under fault attacks and highlights the need to evaluate lightweight cipher designs using nibble-level fault models, along with the traditional bit- and word-level analyses.

This study investigates and improves the differential cryptanalysis of the ARADI block cipher, a low-latency cryptographic system developed by the United States National Security Agency for memory encryption. The preliminary security assessment of ARADI revealed deficiencies that require additional examination. This study revisits and corrects the key recovery attack introduced by Bellini et al. (Indocrypt 2024). By correcting these mistakes, the research presents a precise 11-round key recovery method with revised attack complexities. Additionally, the analysis is expanded to introduce a key recovery attack for 12 rounds of ARADI, leaving the security margin of the cipher to only 4 rounds.

We propose a novel timestamping mechanism for consensus protocols that reliably assigns submission times to honest transactions while preventing adversarial transactions from forging their timestamps. Our mechanism remains secure even under asynchronous networks and in the presence of corrupted parties. We demonstrate how it can be integrated into the three main families of consensus protocols and show its compatibility with layer-2 solutions. This construction enables robust layer-2 implementations that do not rely on timing assumptions, remain resilient under network saturation, and prevent collusion between users and validators to censor honest users.

Random beacons play a critical role in blockchain protocols by providing publicly verifiable, unpredictable randomness essential for secure assignment of protocol roles such as block producers and committee membership. In the interest of efficiency, many deployed blockchains adopt beacon algorithms that suffer from grinding: an adversarial attack in which a party exploits freedom given by the protocol to bias the outcome of the random beacon by resampling it several times and picking the most desirable outcome. To compound the problem, beacons often operate in an iterative manner, where the beacon output produced during one protocol epoch serves as the random seed for the beacon’s invocation in the next epoch. This amplifies the security threat, as such attacks may then aggregate their power over many epochs. In this article, we formulate a generic framework for information-theoretic analysis of grinding in iterated randomness beacons. We define the natural grinding capacity of a beacon, intuitively corresponding to the amount of grinding it allows with a uniformly random seed. We then prove that sufficiently strong tail bounds on this quantity can be transformed into a guarantee on smooth min-entropy of the iterated beacon’s output, even conditioned on all past outputs and irrespective of the inner workings of the beacon. Such min-entropy guarantees can immediately be translated into corresponding statements about various applications of the beacon to committee selection, incentives, or underlying protocol security. Our main technical result concerns conventional longest-chain protocols, where we establish that the combinatorial structure of the forest of longest chains can be leveraged to control grinding. Instantiating the generic framework with these grinding upper bounds, we establish that the randomness beacon of the Ouroboros Praos protocol is secure against adversaries controlling up to about 12% of stake—even without any assumptions bounding the adversarial computational power invested into grinding. This is a qualitatively new guarantee for the protocol.

This paper studies the concrete security of BBS signatures (Boneh, Boyen, Shacham, CRYPTO '04; Camenisch and Lysyanskaya, CRYPTO '04), a popular algebraic construction of digital signatures which underlies practical privacy-preserving authentication systems and is undergoing standardization by the W3C and IRTF. Sch\"age (Journal of Cryptology '15) gave a tight standard-model security proof under the $q$-SDH assumption for a less efficient variant of the scheme, called BBS+--here, $q$ is the number of issued signatures. In contrast, the security proof for BBS (Tessaro and Zhu, EUROCRYPT '23), also under the $q$-SDH assumption, is \emph{not} tight. Nonetheless, this recent proof shifted both standardization and industry adoption towards the more efficient BBS, instead of BBS+, and for this reason, it is important to understand whether this tightness gap is inherent. Recent cryptanalysis by Chairattana-Apirom and Tessaro (ASIACRYPT '25) also shows that a tight reduction to $q$-SDH is the best we can hope for. This paper closes this gap in two different ways. On the positive end, we show a novel tight reduction for BBS in the case where each message is signed at most once--this case covers in particular the common practical use case which derandomizes signing. On the negative end, we use a meta-reduction argument to prove that if we allow generating multiple signatures for the same message, then {\em no} algebraic reduction to $q$-SDH (and its variants) can be tight.

In this paper, we present formally verified proofs of the popular KZG Polynomial Commitment Schemes (PCSs), including the security proofs for the properties of correctness, polynomial binding, evaluation binding and hiding. Polynomial commitment schemes have various applications in cryptography and computer science, including verifiable computation, blockchain and cryptocurrencies, secure multi-party computation as well as in the construction of ZK-SNARKs. To validate security, we utilise EasyCrypt, an interactive theorem prover that allows for formal verification of cryptographic primitives and protocols. This approach enforces correct proofs which cover all required cases and formalising assumptions reducing the risk of overlooked vulnerabilities. This formalisation validates the current understanding of KZG's PCSs as secure while clarifying various issues in the original claims.

The increasing use of multi-party computation (MPC) has spurred the design of symmetric key primitives specifically suited for MPC environments. Recently, weak pseudorandom functions (wPRFs) based on the alternating moduli paradigm have been proposed as a promising class of MPC-friendly primitives. The wPRF proposed at CRYPTO 2024, in its One-to-One parameter set, has been shown to be vulnerable to a key recovery attack dubbed Zeroed-Out, exploiting collisions in the queries. In this paper, we identify that the aforementioned wPRFs conform to a specific structure, called pointwise-keyed function, and further show a different, general key recovery attack. This method, applied to wPRFs in the One-to-One parameter set attacked by Zeroed-out, improves upon the complexity and achieves an attack with complexity below the birthday bound, and stays effective against the proposed countermeasures. For the first time, it succeeds in attacking one of the two Many-to-One parameter sets and stays effective against one of the proposed countermeasures. We also consider its applicability to the alternative wPRF of similar structure proposed by Boneh et al at TCC 2018.

The machine learning problem of model extraction was first introduced in 1991 and gained prominence as a cryptanalytic challenge starting with Crypto 2020. For over three decades, research in this field has primarily focused on ReLU-based neural networks. In this work, we take the first step towards the cryptanalytic extraction of PReLU neural networks, which employ more complex nonlinear activation functions than their ReLU counterparts. We propose a raw output-based parameter recovery attack for PReLU networks and extend it to more restrictive scenarios where only the top-m probability scores are accessible. Our attacks are rigorously evaluated through end-to-end experiments on diverse PReLU neural networks, including models trained on the MNIST dataset. To the best of our knowledge, this is the first practical demonstration of PReLU neural network extraction across three distinct attack scenarios.

We present cryptographic personas, an approach for facilitating access to pseudonymous speech within a community without enabling abuse. In systems equipped with cryptographic personas, users are able to authenticate to the service provider under new, unlinkable personas at will and post messages under those personas. When users violate community norms, their ability to post anonymously can be revoked. Cryptographic personas draws on techniques from recent work, zk-promises. We develop two significant improvements over existing work that make it possible to integrate cryptographic personas into real-time applications like group messaging: we show how to push expensive proof generation into an offline phase and find a way to optimize server-side overhead using recent proof folding techniques. We implement cryptographic personas, integrating them into Signal and Slack, and show that they are concretely efficient enough for deployment.

AI-powered attacks on Learning with Errors (LWE), an important hard math problem in post-quantum cryptography, rival or outperform "classical" attacks on LWE under certain parameter settings. Despite the promise of this approach, a dearth of accessible data limits AI practitioners' ability to study and improve these attacks. Creating LWE data for AI model training is time- and compute-intensive and requires significant domain expertise. To fill this gap and accelerate AI research on LWE attacks, we propose the TAPAS datasets, a Toolkit for Analysis of Post-quantum cryptography using AI Systems. These datasets cover several LWE settings and can be used off-the-shelf by AI practitioners to prototype new approaches to cracking LWE. This work documents TAPAS dataset creation, establishes attack performance baselines, and lays out directions for future work.

Permutation and multiset checks underpin many SNARKs, yet existing techniques either incur superlinear prover time or rely on auxiliary commitments with soundness error that grows linearly in the input size. We present new arguments with linear-time provers and logarithmic soundness, without auxiliary commitments. Prior work achieving logarithmic soundness error arithmetizes the permutation as a product of several multilinear polynomials, a formulation chosen for compatibility with the classic Sumcheck PIOP. A simpler alternative treats permutations as multilinear extensions of their permutation matrices. While this formulation was previously believed to require quadratic prover time, we show that this overhead can be eliminated by taking a linear-algebraic perspective. This viewpoint has a key advantage: partially evaluating the multilinear polynomial of the permutation requires no additional field operations and amounts to applying the inverse permutation to the verifier's challenge vector. This makes the step essentially free in terms of algebraic cost, unlike in prior approaches. Compared to concurrent work BiPerm (Bünz et al., ePrint Archive, 2025), our scheme requires no permutation preprocessing and supports prover-supplied permutations. We show a sparsity-aware PCS like Dory (Lee, TCC, 2021) can compile our PIOP to a SNARK such that the resulting SNARK prover still runs in time $O(n)$. Our construction is the first logarithmically-sound SNARK with an $O(n)$-time prover for both permutation and multiset checks. We further prove a matching optimal prover lower bound, and we identify specific permutations that can be evaluated by the verifier in $O(\mathrm{polylog}(n))$-time. The ability to evaluate these permutations in $O(\mathrm{polylog}(n))$ time allows the verifier to avoid relying on prover-supplied commitments or evaluation proofs. As a result, we obtain the first logarithmically sound, field-agnostic SNARK with an $O(n)$-time prover in this setting.

HQC (Hamming Quasi-Cyclic) was selected as the fifth algorithm in the NIST suite of post-quantum cryptographic (PQC) standards. As the only code-based algorithm currently standardized by NIST, HQC offers a good balance between security assurance, performance, and implementation simplicity. Most existing power analyses against HQC are of the SPA style: they can recover secrets with a small number of traces, but can only tolerate limited noise. In this paper, we develop a chosen-ciphertext DPA-style attack methodology against HQC. We formalize a dedicated chosen-ciphertext setting in which the adversary selects $(\mathbf{u},\mathbf{v})$ to target the intermediate value $\mathbf{v}\oplus(\mathbf{u}\mathbf{y})$ over $\mathbb{F}_2[x]/(x^n-1)$. We further optimize the attack by reducing its computational complexity and generalizing it to target masked HQC implementations. The proposed approach is validated through both simulation and practical experiments. In noiseless simulations, full-key recovery is achieved with just \(10\) traces, and the required number of traces increases linearly with 1/SNR. In practical evaluations on an STM32F4 microcontroller, the secret key can be recovered with \(50\) traces without profiling and \(20\) traces with profiling. When first-order masking is applied, key recovery on the same hardware target remains feasible by exploiting second-order features, requiring approximately \(3{,}000\) traces without profiling. Our results establish a direct and analyzable connection between leakage on \(\mathbf{v}\oplus (\mathbf{u}\mathbf{y})\) and end-to-end key recovery, emphasizing the necessity of higher-order masking countermeasures for HQC implementations.

Privacy-oriented cryptocurrencies like Zerocash only support direct payments and not the execution of more complex contracts. Bitcoin and Ethereum, on the other hand, cannot guarantee privacy, and using them for contract execution leaves open questions about the fungibility of proceeds and requires contract designers to take frontrunning countermeasures. This work reconciles the two worlds and develops a lightweight protocol for decentralized execution of complex contracts that (1) is undetectable to the network at large, (2) maintains anonymity of the potentially mutually distrustful counterparties and unlinkability of their repeated interactions, (3) guarantees fair termination, and (4) is immediately resistant to frontrunning and miner bribery attacks. This is achieved by leveraging the confidentiality and anonymity guarantees of Zerocash and the verifiability and flexibility of trusted execution environments.

Existing PVSS schemes suffer from at least $O(n)$ online complexity due to the need to individually encrypt and prove/ verify each of the $n$ shares. In this work, we present a generic framework for constructing PVSS schemes with $O(1)$ complexity for share distribution and (the expected to be repeated numerous times) public verification. Our key insight lies in establishing a novel connection between PVSS and CCA2-Secure threshold encryption (CCATE), which enables public verifiability enforced by Non-Interactive Zero-Knowledge (NIZK) proofs. We show that a CCATE scheme can be generically transformed into a secure PVSS scheme, eliminating the $O(n)$ bottleneck per on-line operations. We instantiate the framework by presenting two CCATE constructions: 1) A pairing-free scheme based on a committee-based Distributed Key Generation (DKG) protocol and Threshold ElGamal encryption. 2) A silent setup scheme leveraging a non-interactive distributed key generation, relying on Power-of-Tau ceremony. Furthermore, we introduce solutions for dynamic membership updates in both DKG constructions, demonstrating their practicality and adaptability for real-world applications. The scheme is based on an off-line setup stage (before a specific value to share is given) where the $O(n)$ complexity is dealt with. Although our schemes incur higher setup costs, they drastically reduce the complexity of the critical distribution and verification stages to constant time. This trade-off marks a significant advancement in the scalability of PVSS-based systems, especially in the context of blockchain modern transactions. Conceptually, the work points out how variants of the notion of Threshold Encryption can potentially serve as a ``compression mechanism'' for information sharing schemes.

Germany is currently rolling out an opt-out, nation-scale database of the medical records of the majority of its population, with low-income people being disproportionally represented among its users. While there has been considerable criticism of the system coming from civil society, independent academic analysis of the system by the cryptography and information security community has been largely absent. In this paper, we aim to raise awareness of the system’s existence and, based on the system’s public specifications, highlight several concerning cryptographic engineering decisions. Our core observations is that the system’s most sensitive long-term user keys are derived by a rudimentary, home-grown centralized key escrow mechanism. This mechanism relies on a per-use salt and only 256 bit of entropy, shared globally across millions of users. Furthermore, the system’s specification mandates only level 3 compliance with the obsolete FIPS 140-2 security standard, which requires “hard, opaque potting”, but lacks active tamper sensing. As a result, the system remains vulnerable to attacks by nation states and other well-funded adversaries.

Security Meshes are patterns of sensing traces covering an area that are used in Hardware Security Modules (HSMs) and other systems to detect attempts to physically intrude into the device's protective shell. State-of-the-art solutions manufacture meshes in bespoke processes from carefully chosen materials, which is expensive and makes replication challenging. Additionally, state-of-the-art monitoring circuits sacrifice either monitoring precision or cost efficiency. In this paper, we present an embeddable security mesh monitoring circuit constructed from low-cost, standard components that utilizes Time Domain Reflectometry (TDR) to create a unique fingerprint of a mesh. Our approach is both low-cost and precise, and enables the use of inexpensive standard Printed Circuit Boards (PCBs) as security mesh material. We demonstrate a working prototype of our TDR circuit costing less than 10 € in components that achieves both time resolution and rise time better than 200 ps—a 25 × improvement over previous work. We demonstrate a simple classifier that detects several types of advanced attacks such as probing using an oscilloscope probe or micro-soldering attacks with no false negatives.

In this paper, we analyze the clash between privacy-oriented cryptocurrencies and emerging legal frameworks for combating financial crime, focusing in particular on the recent European Union regulations. We analyze Monero, a leading "privacy coin" and a major point of concern for law enforcement, and study the scope of due diligence that must be exercised under the new law with regard to Monero trading platforms and how it translates to the technical capabilities of the Monero protocol. We both recognize flaws in the legislation and identify technical pitfalls threatening either the effective compliance of, say, Monero exchanges or the anonymity endeavour of Monero itself. Of independent interest is that we turn to anamorphic cryptography (marking one of the first practical applications of the concept) and leverage it to build a hidden transaction layer embedded in the Monero blockchain that obfuscates illegal money flow and circumvents transaction-level attempts at enforcing the EU law.

Over the past decades, extensive research has been conducted on lightweight cryptographic primitives. The linear layer plays an important role in their security. In this paper, we propose a family of linear layers consisting of XORs and rotations, which is called multiple rows mixers (MRM). It is a family designed for LS-type ciphers, but mixing elements from several rows. We investigate the impact of the linear layers on the 3-round trail weight of permutations and explore the properties of the inverse of the linear layers with a low XOR count. We employ a generic and extensible approach to determine the parameters of MRM. This approach can automatically generate linear layers that meet the requirements of a given branch number. By applying these design principles and methods, we derive a linear layer that has a dimension of 5 × 64, a differential branch number of 12, a linear branch number of 5 and a computational cost of 2.6 XOR operations per bit. MRM is not limited to fixed dimension and can be extended to other dimensions. In addition, we present a concrete instantiation of a 320-bit permutation using a more efficient instance of MRM, named Hsilu. Its non-linear layer employs the χ operating on columns. Compared with the permutations of Gaston and NIST lightweight standard Ascon, the round function of Hsilu requires fewer XOR operations. Hsilu exhibits competitive security and performance with Ascon and Gaston. We demonstrate that the best-found 3-round differential and linear trails of Hsilu have much higher weights than those of Ascon. Hsilu outperforms Gaston and Ascon in terms of both software and hardware performance.

Private Simultaneous Messages (PSM) and Conditional Disclosure of Secrets (CDS) are two fundamental primitives in information-theoretic cryptography. In a PSM protocol, each party sends a single message to a referee, who can then compute a function $f$ of their private inputs but learns nothing else. A CDS protocol follows the same model as PSM, but the goal is to disclose a secret shared among all parties to the referee if and only if the function $f$ evaluates to $1$. Minimizing the communication complexity of these primitives is a central question in this area. Since the best known constructions for general functions require high communication complexity, recent studies have attempted to obtain more efficient constructions by focusing on symmetric functions, whose outputs are invariant under permutations of inputs. However, the extent to which exploiting symmetry can improve efficiency has remained unclear. In this work, we show upper and lower bounds that relate the optimal communication complexities of PSM and CDS for symmetric functions to those for general functions. When the number of parties is larger than the input domain size, our upper bound for PSM improves the best known communication complexity for symmetric functions. Furthermore, our upper bound for CDS demonstrates for the first time that symmetry can be exploited to reduce communication in CDS protocols. In contrast, when the number of parties is constant, our lower bounds show that focusing on symmetric functions yields only a constant-factor improvement. We also derive an analogous implication for PSM protocols under a plausible conjecture. In addition, our new constructions for symmetric functions lead to improvements over the state-of-the-art results in related models such as ad hoc PSM and secret sharing.

We present a simple IND-CCA lattice-based threshold KEM. At a high level, our design is based on the BCHK transform (Canetti et al., EUROCRYPT 2004), which we adapt to the lattice setting by combining it with the FO transform (Fujisaki and Okamoto, PKC 1999) in order to achieve decryption consistency. As for the BCHK transform, our construction requires a threshold identity-based encryption (TIBE) scheme with suitable properties. We build such an IBE by combining the ABB IBE (Agrawal, Boneh, Boyen, EUROCRYPT 2010) with recent advances in lattice threshold cryptography, such as the threshold-friendly signature Plover (Esgin et al., EUROCRYPT 2024) and a variant of the Threshold Raccoon scheme (Katsumata et al., CRYPTO 2024). The security proof of our scheme relies on a new assumption which we call the Coset-Hint-MLWE assumption, and which is a natural generalisation of the Hint-MLWE assumption (Kim et al., CRYPTO 2023). We prove the hardness of Coset-Hint-MLWE under standard assumptions. We believe this new assumption may be of independent interest. Unlike prior works on IND-CCA lattice-based threshold KEMs, our construction only relies on simple algorithmic tools and does not use heavy machinery such as multi-party computation or threshold fully homomorphic encryption.

Encrypted matrix multiplication (MM) is a fundamental primitive in privacy-preserving machine learning and encrypted data search, but it remains a significant performance bottleneck. Recently, Bae et al.~(Crypto’24) and Park~(Eurocrypt’25) introduced novel algorithms for ciphertext–plaintext (CPMM) and ciphertext–ciphertext (CCMM) matrix multiplications. These algorithms reduce encrypted MM operations to plaintext matrix multiplications (PPMM), enabling implementation through highly optimized BLAS libraries. While these reduction-based methods offer significant improvements, their benefit is limited to scenarios where the matrix dimension $d$ is comparable to the ring dimension $N$ in RLWE-based CKKS schemes. As a result, they fall short for matrix multiplications involving small or medium-sized matrices. We extend the reduction-based CPMM/CCMM into small-sized matrix operations by batching instances. We encode a batch of matrices into a single matrix over algebraic integers, where each entry is obtained by applying the inverse Discrete Fourier Transform to the batch matrix entries at the same position. This encoding enables reductions of encrypted batch MM algorithms to a small number of batch PPMMs, which can be efficiently accelerated by BLAS libraries. Our batch encrypted MM flexibly accommodates diverse matrix dimensions and batch sizes independent of the ring dimension $N$, thereby extending its applicability to practical real-world settings. For two $d \times d$ matrices with $N/d$ batches, our batch CPMM and CCMM algorithms achieve $O(d^2N)$ cost, improving over Bae et al.'s $O(dN^2)$ and Jiang et al.'s $O(d^2N\log N)$~(CCS'18). We further extend our techniques to rectangular matrices, achieving $O(dN^2)$ for multiplying a $d \times N$ matrix by an $N \times N$ matrix, improving previous $O(N^3)$ methods. A proof-of-concept implementation validates these improvements: multiplying 128 batches of $64 \times 64$ matrices takes $0.20$s (CPMM) and $1.08$s (CCMM), yielding $240\times$ and $52\times$ speedups over previous methods. For a $64 \times 2048$ by $2048 \times 2048$ multiplication, our CCMM completes in $7.5$s, achieving a $29\times$ speedup compared to Park's algorithm.

The NIST Post-Quantum Cryptography (PQC) standardization has entered its fourth round, underscoring the critical importance of addressing side-channel attacks (SCA), a dominant threat in real-world cryptographic implementations, especially on embedded devices. This paper presents a novel chosen-ciphertext side-channel attack against CRYSTALS-Kyber (standardized as ML-KEM) implementations with Fisher-Yates shuffled polynomial reduction. We propose an efficient and fault-tolerant key recovery algorithm that, by crafting malicious ciphertexts, induces changes in the Hamming weight distribution of an intermediate polynomial's coefficients (the output of the shuffled polynomial reduction during decapsulation), enabling recovery of secret key coefficients from these changes. To ensure robustness, we propose an error-correction strategy that leverages the Hamming weight classifier's behavior to constrain and shrink the correction search space, maintaining effectiveness even with less accurate classifiers or in low-SNR environments. A Multi-Layer Perceptron (MLP) is employed for Hamming weight classification from side-channel traces, achieving 97.11% accuracy. We combine statistical analysis with explainable deep learning for precise trace segmentation during pre-processing. Experimental results demonstrate full key recovery with only an average of $10 + 354 \times 3$ ciphertext queries and a success rate of 97.98%, reducing the adversarial effort by 95.36% compared to contemporary bit-flip techniques. Although shuffling aims to disrupt temporal correlations, our results show that statistical features persist and leak through shuffled implementations. This work reveals enduring SCA risks in shuffled implementations and informs a broader reassessment of PQC side-channel resilience.

Aggregate signatures allow compressing multiple single-signer signatures into a single short aggregate signature. This primitive has attracted new attention due to applications in blockchains and cryptocurrencies. In multisig addresses, which is one of such applications, aggregate signatures reduce the sizes of transactions from multisig addresses. Security of aggregate signatures under adaptive corruptions of signing keys is important, since one of the motivations of multisig addresses was a countermeasure against signing key exposures. We propose the first aggregate signature scheme tightly secure under adaptive corruptions using pairings. An aggregate signature includes two source group elements of bilinear groups plus a bit vector whose length is equal to the number of single-signer signatures being aggregated. To construct a scheme, we employ a technique from quasi-adaptive non-interactive zero-knowledge arguments. Our construction can be seen as modularization and tightness improvement of Libert et al.'s threshold signature scheme supporting signature aggregation (Theoretical Computer Science 645) in a non-threshold setting.

Side-channel analysis (SCA) exploits physical leakages such as power consumption or electromagnetic emissions to extract secret information from cryptographic implementations. Leakage model is the critical link between observable physical emanations (e.g., power consumption) and internal cryptographic states. When a leakage model fails to match the device’s underlying physical leakage, even powerful attacks like Correlation Power Analysis (CPA) are rendered ineffective. This fundamental challenge limits the success of traditional SCA, especially against noisy or masked targets. In this work, we introduce the Neural Leakage Model (NLM), a novel neural network architecture trained to learn and accurately characterize complex physical leakage to enhance CPA. Moving beyond NLM's superior attack performance, we directly address the critical ``black box'' problem in deep learning-based side-channel analysis (DLSCA) by proposing a novel methodology to transform the trained NLM into a mathematically equivalent, closed-form polynomial expression. This provides interpretability, offering evaluators transparent insight into the precise leakage model the network has learned. We validate NLM’s performance across four diverse datasets spanning three platforms, from low-end microcontrollers to high-end multi-core ARM systems and dedicated FPGAs. Notably, NLM successfully recovers the secret key from the highly challenging and noisy CHES2025 dataset, representing the first known successful CPA attack on this benchmark. Furthermore, NLM demonstrates superior or comparable performance when extended to higher-order CPA against masking countermeasures. Our findings establish NLM as a powerful, generalizable, and interpretable approach to modern side-channel analysis.

Very recently, Crites et al. (CRYPTO 2025) gave a proof for the full adaptive security of FROST (Komlo and Goldberg, SAC 2020), the state-of-the-art two-round threshold Schnorr signature scheme, which is currently used in real-world applications and is covered by an RFC standard. Their security proof, however, relies on the computational hardness of a new search problem they call “low-dimensional vector representation” (LDVR). In fact, the authors show that hardness of LDVR is necessary for adaptive security of a large class of threshold Schnorr signatures to hold, including FROST and its two-round variants. Given that LDVR is a new assumption and its hardness has not been seriously scrutinized, it remains an open problem whether a two-round threshold Schnorr signature with full adaptive security can be constructed based on more well-established assumptions. In this paper, we resolve this open problem by presenting ms-FROST. Our scheme is partially non-interactive and supports any t - 1 < n adaptive corruptions, where n is the number of signers and t is the signing threshold. Its security relies on the algebraic one-more discrete logarithm (AOMDL) assumption, the algebraic group model (AGM), and the random oracle model (ROM). Further, it achieves the strongest security notion (TS-UF-4) in the security hierarchy of Bellare et al. (CRYPTO 2022). To justify our use of the algebraic group model, we show an impossibility result: We rule out any black-box algebraic security reduction in the ROM from AOMDL to the adaptive TS-UF-0 security of ms-FROST.

Keyword Private Information Retrieval (KPIR) enables clients to retrieve database entries by keyword without revealing their query, with many efficient schemes relying on Fully Homomorphic Encryption (FHE). However, existing FHE-based KPIR schemes fix parameters to a bounded computation depth, which becomes a bottleneck in privacy-sensitive settings such as distributed ledger systems and private web search, where databases grow dynamically and clients are lightweight: exceeding this bound incurs significant overhead, forces larger parameters, and leaves any post-retrieval computation to the client or to communication-heavy interactive protocols. We introduce Keyword PIR with Computation (KPIR-C), a new paradigm that enables arbitrary, non-interactive server-side computation on query responses while preserving query privacy. Unlike prior designs, KPIR-C fixes parameters independently of database size or computation depth by leveraging bootstrapping via a co-design of Torus Fully Homomorphic Encryption (TFHE) and KPIR that keeps bootstrapping overhead practical. We further introduce Keyword PIR with Default (KPIR-D) to handle mismatched queries non-interactively. We instantiate KPIR-C with two constructions: one based on equality circuits and constant-weight codes, and one based on Learning With Errors (LWE)-based KPIR, offering complementary communication-runtime trade-offs. We demonstrate experimentally that both achieve competitive performance compared to state-of-the-art KPIR schemes that do not support arbitrary computation.

The European Commission's 2022 proposal for a regulation on child sexual abuse material, popularly labelled ChatControl, obliges online services to detect, report, and remove prohibited content, through client-side scanning. This paper examines the proposal as a case of undone science in computer security ethics: a domain where technical feasibility and rights-compatibility questions remain systematically underexplored. Combining legal analysis with philosophy of technology, the paper argues that client-side scanning transforms end-to-end encryption from a right to secrecy into a conditional privilege of use. By integrating Isaiah Berlin's concept of negative liberty, Langdon Winner’s account of the politics of artifacts, and David Hess’s notion of undone science, the analysis traces how design choices become moral constraints. The discussion situates the European debate within broader concerns about proportionality, epistemic selectivity, and the governance of digital infrastructures. Ultimately, the study shows that the controversy over ChatControl is not only about privacy or child protection but about the epistemic norms that define what counts as legitimate technological knowledge.

We present FaFROST+, the first two-round Schnorr threshold signature scheme achieving full adaptive security under the Algebraic One-More Discrete Logarithm (AOMDL) assumption with identifiable aborts. Building on the FROST framework, FaFROST+ preserves its communication efficiency while dispensing with the LDVR assumption introduced at CRYPTO'25. Our result demonstrates that round-optimal adaptive security, practical efficiency, and identifiable aborts are simultaneously achievable under the well-studied AOMDL assumption.

Over time, cryptographically deniable systems have come to be associated in computer-science literature with the idea of "denying" evidence in court — specifically, with the ability to convincingly forge evidence in courtroom scenarios, and relatedly, an inability to authenticate evidence in such contexts. Indeed, in some cryptographic models, the ability to falsify mathematically implies the inability to authenticate. Evidentiary processes in courts, however, have been developed over centuries to account for the reality that evidence has always been forgeable, and relies on factors outside of cryptographic models to seek the truth "as well as possible" while acknowledging that all evidence is imperfect. We argue that deniability does not and need not change this paradigm. Our analysis highlights a gap between technical deniability notions and their application to the real world. There will essentially always be factors outside a cryptographic model that influence perceptions of a message's authenticity, in realistic situations. We propose the broader concept of credibility to capture these factors. The credibility of a system is determined by (1) a threshold of quality that a forgery must pass to be "believable" as an original communication, which varies based on sociotechnical context and threat model, (2) the ease of creating a forgery that passes this threshold, which is also context- and threat-model-dependent, and (3) default system retention policy and retention settings. All three aspects are important for designing secure communication systems for real-world threat models, and some aspects of (2) and (3) may be incorporated directly into technical system design. We hope that our model of credibility will facilitate system design and deployment that addresses threats that are not and cannot be captured by purely technical definitions and existing cryptographic models, and support more nuanced discourse on the strengths and limitations of cryptographic guarantees within specific legal and sociotechnical contexts.

We consider reversely-degraded wiretap channels, for which the secrecy capacity is zero if there is no channel feedback. This work focuses on a seeded modular code design for the Gaussian wiretap channel with channel output feedback, combining universal hash functions for security and learned feedback-based codes for reliability to achieve positive secrecy rates. We study the trade-off between communication reliability and information leakage, illustrating that feedback enables agreeing on a secret key shared between legitimate parties, overcoming the security advantage of the wiretapper. Our findings also motivate code designs for sensing-assisted secure communication, to be used in next-generation integrated sensing and communication methods.

Oblivious Pseudorandom Function (OPRF) protocols can be categorized into two variants: chosen-key OPRFs -- where the server provides the PRF key $k$ as an input -- and ephemeral-key OPRFs -- where the functionality randomly samples $k$ on behalf of the server. Ephemeral-key OPRFs can be constructed from simple cryptography, such as black-box OT and a random oracle. Chosen-key OPRFs, on the other hand, are only known by employing one of the following (more expensive) approaches: - Express the underlying PRF as a circuit, then use generic MPC techniques to evaluate it in a non-black-box manner. - Base the underlying PRF on some specific public-key assumption, such as RSA or DDH, then build a custom protocol that achieves obliviousness by leveraging the PRF's algebraic structure. Thus, there exists a qualitative gap between known instantiations of these two OPRF variants, in terms of both assumptions and efficiency. We show that this gap is inherent. In particular, one might hope for an OPRF with all of the following characteristics: the protocol (1) is chosen-key, (2) supports domains of super-polynomial size, and (3) is constructed from "simple" cryptography, such as the minimal assumption of OT, plus a random oracle. We show that no such protocol can exist. Let $\Pi$ be any chosen-key OPRF protocol defined in terms of a PRF $F$, where $F$ is defined with respect to (only) a random oracle. This restriction on $F$ rules out the above approaches of either using a PRF in a non-black-box way or basing the underlying PRF itself on public-key cryptography. While $F$ is restricted in its use of cryptography, the protocol $\Pi$ is not: $\Pi$ may use arbitrary cryptography, e.g., OT, FHE, iO, etc. We show that each invocation of any such $\Pi$ necessarily leaks information about the server's key $k$. After a bounded number of queries, an adversarial client can effectively recover $k$, breaking server privacy. To complement our negative result, we provide a matching positive result: we construct a chosen-key OPRF from black-box OT and RO, where server privacy holds for some bounded number of queries $n$. This protocol's underlying PRF is constructed from a $(n+1)$-wise independent hash function and RO; the server's key $k$ has length scaling linearly in $n$ which, by our lower bound, is optimal. Thus, our two results tightly (i.e., up to $\mathrm{poly}(\lambda)$ factors) characterize (im)possibility for chosen-key OPRFs, unless one uses non-black-box cryptography or a public-key-style PRF.

Distributed Key Generation (DKG) is essential for secure, decentralized cryptographic systems, enabling collaborative key pair generation without a trusted authority. This capability underpins critical applications such as threshold signatures and blockchain-based protocols. To achieve post-quantum security, existing robust lattice-based DKG protocols, tailored for synchronous networks, rely on complaint-based Verifiable Secret Sharing (VSS). However, these protocols lack public verifiability and compatibility with asynchronous environments, constraining their use in Byzantine fault-tolerant settings. This paper presents LADKG, the first Lattice-Based Asynchronous Distributed Key Generation protocol designed for post-quantum secure and scalable distributed systems. LADKG integrates Asynchronous Verifiable Short Secret Sharing (AV3S) with an Approximate Asynchronous Common Subset (AACS) protocol to achieve efficient key generation. By deferring verification and leveraging deterministic approximate agreement, LADKG reduces computational and communication overhead while maintaining security and robustness. Evaluations on geo-distributed AWS EC2 clusters demonstrate that LADKG is comparable or better than classical Asynchronous Distributed Key Generation (ADKG) schemes in scalability and efficiency. Under optimistic conditions with $n=121$ nodes, completion is achieved in 45 seconds, ensuring robust key generation for post-quantum secure applications.

In this note we identify major issues with “Exact Coset Sampling for Quantum Lattice Algorithms” by Zhang. The issues include: • in the first version, the proposed algorithm requires foreknowledge of the answer to compute the answer; • in the latest version, the proposed algorithm displays basic misunderstandings of quantum mechanics. We believe the existence of these issues refute the author’s claims.

Software masking---particularly through threshold implementations---has long been regarded as a foundational defense mechanism against side-channel attacks. These schemes enforce the principles of non-completeness and uniformity, offering provable first-order resistance even under realistic leakage assumptions. However, such assurances were primarily developed under the simplified assumption of scalar or in-order execution, where instruction flow and data dependencies are well-behaved and predictable. Contemporary processors, by contrast, employ deeply optimized out-of-order (OoO) pipelines that rely on dynamic scheduling, register renaming, operand forwarding, and speculative execution. These aggressive microarchitectural behaviors, while improving performance, also create intricate dataflow interactions that can inadvertently violate the independence of masked shares. As a result, secret shares that are theoretically isolated may recombine transiently through register reuse, speculative reordering, or transition-based leakage in the physical register file. This raises a critical question: to what extent do masking countermeasures remain secure when executed on modern OoO architectures? To explore this question, we develop a systematic methodology for architectural leakage analysis based on fine-grained execution traces from an OoO RISC-V processor. The analysis correlates microarchitectural events with instruction-level behavior, reconstructs the evolution of physical-register assignments, and identifies instances where fundamental masking principles such as non-completeness or uniformity are violated. Through this structured approach, we establish a taxonomy of microarchitectural leakage classes and formalize two dominant categories: (1) \textit{Rename-Induced Transition Leakage (RIL)}, arising from register reuse and transitional dependencies between masked shares; and (2) \textit{IEW-Induced Dispatch Leakage}, originating from speculative instruction issue and execution reordering. The methodology is demonstrated across three representative studies: (i) a masked Toffoli gate, where violations of non-completeness are revealed; (ii) a masked PRESENT cipher, where 571 leakage events are identified due to register-induced transitions; and (iii) a Probe Isolating Non-Interference (PINI) experiment, which exposes how speculative and dynamic scheduling can compromise isolation guarantees, thereby weakening the composability of masked implementations that are provably secure under idealized models. These results underline a fundamental insight: the security of masking countermeasures cannot be evaluated in abstraction from the hardware on which they execute. Ensuring true resistance to side-channel attacks demands a holistic view that jointly considers both the algorithmic soundness of masking constructions and the microarchitectural realities of modern processors.

Batch arguments for NP (BARGs) are non-interactive proof systems that allow a prover to convince a verifier that $k$ NP statements $x_1, \ldots, x_k$ are valid relative to some circuit $C$, i.e. there exist witnesses $w_i$ such that $(x_i, w_i)$ satisfy $C$ for all $i$, while the proof size remains sublinear in $k$. Most existing BARG constructions achieve a proof size of $|\pi| = poly(\lambda, |C|, \log k)$ for large or not explicitly specified $poly$ acting on $|C|$, with two exceptions: - Devadas et al. and Paneth and Pass's ''rate-1'' constructions [FOCS'22] achieve $|\pi| = |w| + O(|w|/\lambda) + poly(\lambda, \log k)$ (with matching verification time for Paneth and Pass), but for not explicitly specified $poly$ due to non-black-box use of cryptographic primitives. - Waters and Wu's algebraic (pairing-based) construction [Crypto'22] and follow-up works achieve $|\pi| = O(\lambda \cdot |C|)$. In this work, we give the first algebraic (pairing-based) construction of BARG that achieves proof size and online verifier runtime $O(\lambda \cdot |w|)$. We achieve our result by means of a compiler which builds a BARG generically from a projective chainable functional commitment (PCFC), which supports somewhere extraction, subvector projection, and functional openings. We then construct a PCFC from the standard MDDH assumption in bilinear groups by building on top of the functional commitment for circuits by Wee and Wu [Eurocrypt'24]. Our black-box transformation may be of independent interest for understanding the connection between functional commitments and BARGs and towards obtaining other algebraic constructions of the latter.

Homomorphic Encryption (HE) allows parties to securely outsource data while enabling computation on encrypted data, protect- ing against malicious parties and data leakages. More recent HE schemes enable approximate arithmetic on complex vectors and approximation of non-linear functions, specifically useful for image processing algorithms. The Fourier Shape Descriptor (FSD) is a classical method for shape matching via frequency-domain representation, and we show that FSD can be computed entirely in the encrypted domain. To the best of our knowledge, this is the first work to implement secure shape descriptors and matching via HE. We also present two experiments for similar and different shapes, and measure the performance of the encrypted algo- rithm.

Threshold signatures are an important tool for trust distribution, and preserving the interface of standardized signatures, such as Schnorr signatures, is crucial for their adoption. In practice, latency dominates end-to-end signing time, so minimizing the number of interaction rounds is critical. Ideally, this is achieved under minimal assumptions and with adaptive security, where the adversary can corrupt signers on-the-fly during the protocol. While Schnorr signatures are proven secure under the Discrete Logarithm (DL) assumption in the random oracle model, the most round-efficient adaptively-secure threshold Schnorr protocols rely on stronger assumptions such as Decisional Diffie-Hellman (DDH), the Algebraic One-More Discrete Logarithm (AOMDL) or even the Low-Dimensional Vector Representation (LDVR) assumptions. The only adaptively-secure threshold Schnorr signature from the DL assumption requires five rounds, leaving a significant gap in our understanding of this fundamental primitive. Achieving low-round protocols with adaptive security from the DL assumption has remained an open question. We resolve this question by presenting the first adaptively-secure threshold Schnorr scheme in three rounds (two online, one offline) in the random oracle model under the DL assumption. Our result demonstrates that achieving both low round complexity and adaptive security is possible while preserving the (so far) minimal assumptions for Schnorr signatures. To achieve this, we introduce new techniques, including a novel use of an equivocal commitment scheme paired with a simulation-extractable NIZK, and a masking-based aggregated opening strategy for homomorphic commitments. Our work also makes several contributions that might be of independent interest, including a formalization of a strong adaptive security notion, a stronger commitment equivocation property, and an analysis of the simulation-extractability of the randomized Fischlin transformation.

The lattice trapdoor associated with Ajtai's function is the cornerstone of many lattice-based cryptosystems. The current provably secure trapdoor framework, known as the GPV framework, uses a strong smoothness condition, i.e. $\epsilon\ll \frac{1}{n^2}$ for smoothing parameter $\eta_{\epsilon}(\mathbb{Z}^{n})$, to ensure the correctness of the security reduction. In this work, we investigate the feasibility of weak smoothness, e.g. $\epsilon = O,(\frac{1}{n})$ or even $O(1)$ in the GPV framework and present several positive results. First, we provide a theoretical security proof for GPV with weak smoothness under a new assumption. Then, we present Gaussian samplers that are compatible with the weak smoothness condition. As direct applications, we present several practical GPV signature instantiations based on a weak smoothness condition. Our first instantiation is a variant of Falcon, called, Falcon$^{a ws}$, achieving smaller size and higher security. The public key sizes are $21\%$ to $28\%$ smaller, and the signature sizes are $23.5\%$ to $29\%$ smaller than Falcon. We also showcase an NTRU-based GPV signature scheme that employs the Peikert sampler with weak smoothness. This offers a simple implementation while the security level is greatly lower. Nevertheless, at the NIST-3 security level, our scheme achieves a $49\%$ reduction in size compared to Dilithium-3. We also derive a weak smoothness variant of the Antrag signature scheme, called Antrag$^{ws}$, along with a floating-point free version, which offers an attractive trade-off among portability, efficiency and security. Compared to Antrag, Antrag$^{ws}$ achieves reductions of $16.5\%$ to $21.8\%$ in signature size and $14.3\%$ to $21.4\%$ in public key size. Furthermore, we adapt the compact gadget framework to the weak smoothness setting, removing the need for the new assumption and improving concrete parameters of gadget-based signatures, e.g. Hufu.

In this paper, we propose the Hybrid FAFFT-CRT method for accelerating HQC polynomial multiplication on the ARM Cortex-M4. The method uses the Chinese Remainder Theorem to map the polynomial ring to a product of an FAFFT-friendly ring of size $2^{d+1}$ and a small-degree residual ring, reducing the FAFFT transform length by half compared to the state-of-the-art. We also present the Hybrid Karatsuba-FAFFT method as an alternative hybrid based on a 2-way Karatsuba split, and apply radix-16 multiplication to HQC for the first time, with an operation-count cost model for selecting Karatsuba and Toom-Cook combinations. Additionally, we improve the core FAFFT butterfly through shortened XOR sequences, register scheduling, and SWAPMOVE-based bit swaps. On a NUCLEO-L4R5ZI board with a Cortex-M4 microcontroller, the Hybrid FAFFT-CRT method reduces polynomial multiplication cycles by 36.0% and 29.3% for HQC-1 and HQC-3, leading to cycle reductions for key generation, encapsulation, and decapsulation by 25.5%, 26.1%, and 23.5% for HQC-1, and 19.3%, 20.0%, and 19.1% for HQC-3. For HQC-5, our optimized butterfly FAFFT provides a consistent speedup of 1.4-1.6% across all KEM operations.

We are now entering an era where the large-scale deployment of anonymous credentials seems inevitable, driven both by legislation requiring age verification and the desire to distinguish humans from bots in the face of the proliferation of AI-generated content. However, the widespread deployment of anonymous credentials faces the same security and fraud concerns as existing credentials, but without the established techniques for securing them. For non-anonymous credentials on the web today, authentication is a continuous process in which servers collect large volumes of behavioral data to protect account holders (e.g., by detecting account compromise) or to combat fraudulent behavior. In this paper, we propose Continuous Anonymous Authentication (CAA) schemes and give a concrete construction and applications for preventing credential sharing and theft. CAA schemes allow us to move the server-side collection, storage, and processing of these behavioral signals to the client while maintaining privacy and integrity. CAA schemes support, on the client side, a number of common behavioral analysis tests and analytics both for determining fraudulent behavior and updating security policies. We implement a prototype, zk-Cookies, which runs in the browser, and supports common behavioral signals such as IP address and geolocation history, browser fingerprinting, and page view history. Using this, we build a prototype application for age verification based on legacy credentials (like passports). We implement these checks efficiently in zk-SNARKs, and also show how to securely implement differentially private behavioral analytics in a zk-SNARK. The simplest version of our construction can perform the computation for an update in under 200 ms.

Function Secret Sharing (FSS) schemes enable to share secret functions between multiple parties, with notable applications in anonymous communication and privacy-preserving machine learning. While two-party schemes offer logarithmic key sizes, multi-party schemes remain less practical due to significantly larger keys. Although several approaches have been proposed to improve multi-party schemes, a significant efficiency gap remains between the two-party and multi-party settings. Our work introduces noisy FSS: a relaxation of FSS preserving the standard privacy guarantees but relaxing the correctness definition by allowing a small amount of noise in the output. We formally define noisy FSS and show how the noise introduced by the scheme can be leveraged to provide differential private outputs in statistics applications. To demonstrate the benefits of this relaxation, we adapt a scheme proposed by Corrigan-Gibbs et al. (S&P'15). While their scheme provides the smallest key sizes among multi-party schemes, they do not support some applications notably in statistics due to their non-linear share decoding. On the contrary, recent works such as Goel et al. (CRYPTO'25) have larger keys, but support all FSS applications. Our noisy adapted scheme offers the best of both worlds by matching the best key sizes, while providing the properties necessary to statistics applications.

Masking is one of the most widespread countermeasures against side-channel attacks. However, its integration into hardware implementation is subject to physical hazards that can mitigate its security. To counter glitches, the most studied physical hazard, an effective solution is to resynchronize the signals by integrating additional hardware layers into the architecture. However, these solutions have an impact on the performance of the implementation. A solution to avoid these limitations is to use more shares to compute higher-degree functions. We study the cost of this approach, denominated ($td+n$)-masking. We first derive optimal dependence structures for the creation on non-complete sharings, which allow us to obtain efficient implementation of substitution boxes. As a case study, we use these elements to create a second-order masked architecture for the PRESENT cipher. We perform multiple TVLA tests to validate our approach. Our results confirm that the strategy is efficient in terms of performance, at the cost of increased hardware resources.

We propose an efficient fully homomorphic encryption (FHE) scheme tailored for matrix arithmetic based on the Ring-Learning with Errors (RLWE) problem. The proposed scheme naturally supports matrix multiplication, addition, and Hadamard multiplication for batched matrices of various sizes over both complex numbers and integers. Encrypted matrix multiplication is reduced to four matrix multiplications of ciphertext elements, without the need for expensive operations such as slot-to-coefficient conversion or ring switching. In addition, the scheme efficiently supports matrix transformations, including general and conjugate transpositions, as well as matrix rotations: inter-matrix rotations across batched matrices and intra-matrix rotations within rows and columns. Moreover, the proposed FHE scheme can be directly combined with existing bootstrapping algorithms. By eliminating the need for expensive operations such as repeated slot rotations and conversion between slot- and coefficient-encoding, the proposed construction achieves significant performance improvements. In our construction, encrypted multiplications of $n\times n$ matrices under slot encoding are decomposed into two parts: (1) matrix multiplication — four $n\times n$ matrix multiplications of ciphertext coefficients, and (2) key switching — with a total cost approximately 2–4 times that of Hadamard multiplication. We implemented the proposed scheme and utilized the FLINT library for the matrix multiplication component. Experimental results demonstrate that, even when leveraging highly optimized implementations, matrix multiplication remains the major cost, indicating that our construction substantially reduces auxiliary overheads and achieves strong overall efficiency.

This work presents a RISC-V extension for Post-Quantum Cryptography (PQC) called HORCRUX, which provides a unified Instruction-Set Extension (ISE) supporting all NIST-approved PQC algorithms. HORCRUX addresses the current fragmentation in hardware support, where existing extensions typically focus on individual algorithms or limited subsets of PQC schemes, and targets the common kernels shared across ML-KEM, ML-DSA, SLH-DSA, and HQC. To address the primary computational bottlenecks of all these algorithms (namely, modular arithmetic, matrix multiplication, and hash transformations), the extension introduces new RISC-V instructions executed by a tightly coupled coprocessor. This coprocessor requires only minimal hardware resources, making the architecture efficient for constrained devices while still providing substantial acceleration. An experimental evaluation on a Zynq UltraScale+ FPGA demonstrates speedups of up to 3× for lattice and hash-based schemes, and over 30× for code-based schemes, while adding less than 2,900 LUTs and 400 FFs to the design. The extension’s modular structure maintains backward compatibility with standard RISC-V cores, offering a scalable solution for deploying PQC on highly constrained embedded systems.

Multi-key fully homomorphic encryption (MKFHE) extends the capability of fully homomorphic encryption by enabling homomorphic computations on ciphertexts encrypted under different keys. Multi-key blind rotation is the core and most computationally intensive component of MKFHE. The NTRU-based multi-key blind rotation proposed by Xiang et al. (ASIACRYPT 2024) has the potential to achieve smaller key sizes, faster blind rotation, and lower noise compared to its RLWE-based counterpart. However, while the multi-key blind rotation in the RLWE-based scheme proposed by Kwak et al. (PKC 2024) remains compatible with their single-key counterpart, the NTRU-based versions do not. This motivates our work to advance NTRU-based schemes in both efficiency and compatibility. We propose a novel workflow for NTRU-based multi-key blind rotation that achieves compatibility with its single-key counterpart. Our approach significantly reduces both computational and storage complexity compared to the state-of-the-art NTRU-based design, while maintaining a comparable noise magnitude. Building upon this workflow, we further construct two MKFHE schemes for bootstrapping multi-key LWE ciphertexts and multi-key matrix NTRU ciphertexts, both supporting a super-constant number of parties with respect to the ring dimension $N$. Experimental results demonstrate that our method outperforms existing NTRU-based bootstrapping for TFHE-like MKFHE in both computational efficiency and bootstrapping key size. Specifically, our 2-key gate bootstrapping takes only 26ms and requires a bootstrapping key of size 7.34MB, achieving a 3.1× speedup and a 1.9× key size reduction compared to prior NTRU-based works.

Prange's information set algorithm is a decoding algorithm for arbitrary linear codes. It decodes corrupted codewords of any $\mathbb{F}_2$-linear code $C$ of message length $n$ up to relative error rate $O(\log n / n)$ in $\mathsf{poly}(n)$ time. We show that the error rate can be improved to $O((\log n)^2 / n)$, provided: (1) the decoder has access to a polynomial-length advice string that depends on $C$ only, and (2) $C$ is $n^{-\Omega(1)}$-balanced. As a consequence we improve the error tolerance in decoding random linear codes if inefficient preprocessing of the code is allowed. This reveals potential vulnerabilities in cryptographic applications of Learning Noisy Parities with low noise rate. Our main technical result is that the Hamming weight of $Hw$, where $H$ is a random sample of *short dual* codewords, measures the proximity of a word $w$ to the code in the regime of interest. Given such $H$ as advice, our algorithm corrects errors by locally minimizing this measure. We show that for most codes, the error rate tolerated by our decoder is asymptotically optimal among all algorithms whose decision is based on thresholding $Hw$ for an arbitrary polynomial-size advice matrix $H$.

At CHES 2017, Banik et al. proposed a lightweight block cipher GIFT consisting of two versions GIFT-64 and GIFT-128. Recently, there are lots of authenticated encryption schemes that adopt GIFT-128 as their underlying primitive, such as GIFT-COFB and HyENA. To promote a comprehensive perception of the soundness of the designs, we evaluate their security against differential-linear cryptanalysis. For this, automatic tools have been developed to search differential-linear approximation for the ciphers based on S-boxes. With the assistance of the automatic tools, we find 13-round differential-linear approximations for GIFT-COFB and HyENA. Based on the distinguishers, 18-round key-recovery attacks are given for the message processing phase and initialization phase of both ciphers. Moreover, the resistance of GIFT-64/128 against differential-linear cryptanalysis is also evaluated. The 12-round and 17-round differential-linear approximations are found for GIFT-64 and GIFT-128 respectively, which lead to 18-round and 19-round key-recovery attacks respectively. Here, we stress that our attacks do not threaten the security of these ciphers.

The generic group model (GGM) is fundamental for evaluating the feasibility and limitations of group-based cryptosystems. Two prominent versions of the GGM exist in the literature: Shoup's GGM and Maurer's GGM. Zhandry (CRYPTO 2022) points out inherent limitations in Maurer's GGM by demonstrating that several textbook cryptographic primitives, which are provably secure in Shoup's GGM, cannot be proven secure in Maurer's model. In this work, we further investigate Shoup's GGM and identify novel limitations that have been previously overlooked. Specifically, to prevent generic algorithms from generating valid group elements without querying the oracle, the model typically employs sufficiently large encoding lengths. This leads to sparse encodings, a setting referred to as the sparse generic group model (sparse GGM). We emphasize that this sparseness introduces several constraints: --Groups with AE and Black-Box Separation: Shoup's GGM is typically instantiated with elliptic curve groups, which admit admissible encodings (AE)—functions mapping from Z_N to elliptic curve points. We establish a black-box separation, showing that the sparse GGM fails to capture cryptographic groups that are both (1) computational Diffie-Hellman (CDH) secure and (2) compatible with admissible encodings. --Comparison with EC-GGM: We examine the relationship between the sparse GGM and the Elliptic Curve Generic Group Model (EC-GGM) introduced by Groth and Shoup (EUROCRYPT 2022), which inherently yields CDH-secure groups with admissible encodings. Within the framework of indifferentiability, we prove that EC-GGM is strictly stronger than sparse GGM. --Dense Groups and Black-Box Separation: We revisit groups with dense encodings and establish a black-box separation between CDH-secure dense groups and the sparse GGM. --Extension to Bilinear Settings: Our results naturally extend to the sparse Generic Bilinear Group Model (GBM), demonstrating that the aforementioned constraints still hold. In conclusion, our findings indicate that both feasibility and impossibility results in Shoup's GGM should be reinterpreted in a fine-grained manner, encouraging further exploration of cryptographic constructions and black-box separations in EC-GGM or dense GGM.

The asymmetric cryptographic constructions upon on number- theoretic hardness assumptions have become insecure, due to Shor’s quantum algorithm and they will be vulnerable to large scale quantum computers. Hence, the adaption to quantum-resistant cryptosystems is a major task. Digital signatures, being a fundamental primitive in nu- merous applications. Recently, a new approach by Nguyen et al. [9] has claimed post-quantum security by basing the signature algorithm’s se- curity on a variant of the discrete logarithm problem. In this paper, we present a cryptanalysis of this construction and demonstrate a practi- cal forgery attack that allows generating an unlimited number of valid signatures—without access to a signing oracle.

Post-quantum cryptosystems are currently attracting significant research efforts due to the continuous improvements in quantum computing technologies, which led the National Institute of Standards and Technology (NIST) to open standardization competitions to foster proposals and public scrutiny of cryptosystems and digital signatures. Whilst NIST has chosen, after four selection rounds, three digital signature algorithms, it also has opened a new selection process as the chosen candidates were either relying only on lattice-based computationally hard problems, or had unsatisfactory performance figures. In this work, we propose two optimized implementations of the Codes and Restricted Objects Signature Scheme (CROSS) targeting the Cortex-M4 platform. One implementation targets the minimal possible stack size while the other trades some memory space for performance optimization using DSP instructions for some performance critical arithmetic operations. We show that all parameter sets fit within at maximum 24 kB of stack which corresponds to a reduction by a factor of 15 to 45 with respect to the reference implementation. The memory footprint of our implementation, taking the size of the signature also into account, is less than 128 kB. We additionally outline different stack reduction options which allow for a fine grained trade-off between memory footprint and performance of the algorithm. Notably, we also show that our memory optimizations alone have no significant performance impact on the signature verification of CROSS while we even achieve a speed-up factor of up to 1.7 when taking the stack and speed optimizations into account.

Conventional cryptographic algorithms rely on hard mathematical problems to ensure an appropriate level of security. However, with the advent of quantum computing, classical cryptographic algorithms are expected to become vulnerable. For this reason, Post-Quantum Cryptography (PQC) algorithms have emerged as a response, being designed to resist quantum attacks. Most PQC algorithms rely on the Learning With Errors (LWE) problem, where generating pseudo-random controlled errors is crucial. A well-known solution is the use of hash functions followed by error samplers, implemented according to specific error distributions, whose implementation is challenging. This paper provides a proof of concept demonstrating how Approximate Computing (AxC) can be exploited in LWE-based cryptographic algorithms to alleviate this implementation bottleneck. The main idea is to use AxC circuits to run some of the algorithm's operations, introducing the required error for free thanks to the approximation. Our key contribution is demonstrating how AxC techniques can be effectively applied to LWE-based algorithms, highlighting a novel approach to generating and introducing the error. This concept has proven effective in an approximate implementation of the FrodoKEM algorithm, where we achieve a 50.3% reduction in the need for Gaussian sampling. Additionally, we observe a performance improvement of 2.19%, which further supports the feasibility of this approach. Overall, this work introduces and validates a new design direction for LWE-based cryptography through AxC, opening the way for further research.

This article aims to consider batch hashing to elliptic curves. The given kind of hash functions found numerous applications in elliptic curve cryptography. In practice, a hash-to-curve function is often evaluated at a time by the same entity at many different inputs. It turns out that under certain mild conditions simultaneous evaluation can be carried out several times faster than separate ones. In this regard, the article introduces a new class of elliptic curves over finite fields, more appropriate for multiple hashing to them. Moreover, two explicit hashing-friendly Montgomery/twisted Edwards curves (of $\approx 128$ security bits) have been generated: one of CM discriminant $-7$, i.e., a GLV-friendly curve and one of huge CM discriminant, i.e., a CM-secure curve. The new elliptic curves are intentionally covered by so-called Klein's and Bring's curves of geometric genera $3$ and $4$, respectively. The latter are well studied in various algebraic geometry contexts, although they have not yet been (reasonably) applied in cryptography to the author's knowledge. Such a mathematical complication is justified, since conventional curves (from existing standards or of $j$-invariants $0$, $1728$) are seemingly less efficient for batch hashing.

It is a common practice for symmetric-key ciphers is to encode a cryptanalysis problem as an instance of the Mixed Integer Linear Programming (MILP) and then run the instance with an efficient solver. For this purpose, it is essential to model the components in a way that is compatible with the MILP formulation while preserving the characteristics of the cipher. In this work, we look at the problem of efficiently encoding a substitution box (SBox for short). More specifically, we take the evaluation tables, namely, the Difference Distribution Table (DDT), the Linear Approximation Table (LAT), the Division Property Table (DPT) and the Boomerang Connectivity Table (BCT). In the current stage, the only model proposed/used in the literature is to look for the positive samples (i.e., non-zero) entries in the evaluation table and encode for that (this ultimately leads to the minimum number of active SBoxes for the target cipher). In our first contribution, we experiment with the complementary concept of using the negative samples (i.e., treating the zero entries as non-zero entries and vice versa), finally a proper recovery procedure is performed to get back to the original problem. In our second contribution, we observe that the top row and column of each table (which are always taken into consideration by the past researchers) are actually redundant, as those are inherently taken care of through additional constraints at another level. We thus propose a simplified model that removes those row and column. We show the efficacy of our proposals by furnishing results on the evaluation tables on multiple SBoxes. Our proposals often reduce the number of constraints for certain SBox evaluation tables (depending on the properties of the table) from the state-of-the-art results.

We present Golden, a non-interactive Distributed Key Generation (DKG) protocol. Golden achieves public verifiability in a lightweight, non-interactive manner, outputting Shamir secret shares of a field element $\mathsf{sk} \in \mathbb{Z}_p$ to all participants, and a public key $\mathsf{PK}= g^{\mathsf{sk}}$ that is a discrete-logarithm commitment to sk. Golden avoids the overhead of public-key encryption schemes like ElGamal, Pallier, or class groups by using a novel building block: a two-party exponent Verifiable Random Function (two-party eVRF), which may be of independent interest. We present a concretely efficient construction based on the Diffie-Hellman non-interactive key exchange and an efficient zero-knowledge proof. DKG participants use this two-party eVRF in a pairwise manner to generate a one-time pad to encrypt shares, while maintaining public verifiability. Golden can be securely instantiated in the random oracle model over any elliptic curve for which the discrete logarithm assumption holds. Golden drastically improves efficiency, compared to other non-interactive DKGs. For 50 participants, Golden requires only 223 kb bandwidth and 13.5 seconds of total runtime for each participant, in comparison to ElGamal-based non-interactive DKG, which requires 27.8 MB bandwidth and 40.5 seconds runtime per participant.

Fault injection attacks target cryptographic primitives implemented in hardware by deliberately inducing faults, leading to abnormal behavior that can be exploited to extract sensitive information. To mitigate these threats, practical and low-overhead countermeasures, whether at the algorithmic or physical level, are essential. However, the real-world effectiveness of these countermeasures remains uncertain, as it is not always clear whether, or to what extent, their underlying security assumptions hold in practice. Therefore, a thorough evaluation under realistic attack scenarios, including recent techniques such as Electromagnetic Fault Injection (EMFI), is crucial. In this work, we demonstrate the resistance of a protected real-world target chip against EMFI. Specifically, our fabricated 65 nm CMOS ASIC employs concurrent error correction based on the techniques described in Impeccable Circuits II as an algorithm-level countermeasure. At the physical level, the chip integrates multiple coils of varying sizes and positions that serve as sensors for electromagnetic fields. We employ a practical and affordable attack setup featuring a commercial fault-injection probe mounted on an XYZ stage for precise positioning over the ASIC. This setup allows us to investigate the effects of various attack parameters, including the probe’s position, pulse polarity, and voltage level. Our results highlight that the coils serve as a lightweight and effective countermeasure for the practical detection of EMFI attempts. In contrast, for concurrent error correction, a gap between theory and practice emerges: the protection overhead actually makes such designs more susceptible to EMFI in real-world scenarios.

Deniable authentication is a very desirable guarantee for secure messaging — and in particular Off-The-Record messaging: it allows a sender Alice to authenticate her messages while retaining plausible deniability on whether she authenticated any of her messages. While deniable authentication is not achievable in standard Public-Key Infrastructures, Dodis et al. (TCC ’09) have shown that it is achievable in so-called Key-Registration with Knowledge setups. These guarantee: 1. dishonest parties’ key-pairs are honestly generated; and 2. dishonest parties have access to their secret keys. However, these are expensive assumptions, and it is unclear how to securely realize them in practice. In this paper we study the weakest setup requirements that are sufficient to enable deniable authentication. One of our main findings is that deniable authentication does require an expensive — but not necessarily impractical — Key-Registration procedure. Roughly, this procedure must guarantee that if a user’s key registration is successful, then the user can extract by themselves valid secret keys from their interaction with the registration authority. We show this setup is sufficient by capturing it via an appropriate security model and then proving the anonymity of a ring signature based on it. On the other hand, we explain why this setup seems inherently necessary by listing a series of attacks that void deniability as soon as a user manages to successfully register a public key for which it can convincingly claim it does not know a valid (corresponding) secret key. The requirements we identify for key-registration protocols are expensive, but not necessarily impractical. Our second main contribution is showing how to amortize the per-user Key-Registration cost, which brings deniable authentication guarantees a step closer to practicality.

Unclonable cryptography leverages the quantum no-cloning principle to copy-protect cryptographic functionalities. While most existing works address the basic single-copy security, the stronger notion of multi-copy security remains largely unexplored. We introduce a generic compiler that upgrades collusion-resistant unclonable primitives to achieve multi-copy security, assuming only one-way functions. Using this framework, we obtain the first multi-copy secure constructions of public-key quantum money (termed quantum coins), single-decryptor encryption, unclonable encryption, and more. We also introduce an extended notion of quantum coins, called upgradable quantum coins, which allow weak (almost-public) verification under weaker assumptions and can be upgraded to full public verification under stronger assumptions by the bank simply publishing additional classical information. Along the way, we give a generic compiler that upgrades single-copy secure single-decryptor encryption to a collusion-resistant one, assuming the existence of functional encryption, and construct the first multi-challenge secure unclonable encryption scheme, which we believe are of independent interest.

Zero-Knowledge (ZK) protocols rely internally on hash functions for their security arguments. However, the hash functions that are the most efficient in this context differ substantially from e.g. SHA-3: their round function $R$ must enable an efficient arithmetization of its verification. In practice, it means that verifying if $y = R(x)$ involves as little finite field multiplications as possible. In turn, this design requirement implies a greater vulnerability to algebraic attacks. In fact, improvement of those have proved devastating, and imply the need to completely rethink the methods used to ensure security against them. In this paper, we show that it is possible to build a simple yet efficient security argument based on a precise estimate of the so-called “ideal degree” of a system of equations. Furthermore, we show that the increase of this quantity across rounds is tightly connected to the cost of the hash function in two different arithmetizations, namely AIR and R1CS. We precisely quantify this relation by introducing ALgebraic Figures Of Merit (ALFOMs) that capture how efficient a specific primitive (and in fact its round function) are at increasing the security per unit of cost. This new insight allows us to better understand sometimes puzzling performance differences between state-of-the-art hash functions in the R1CS and AIR cases, and to provide a fair and simple comparison of their round functions in this context. Furthermore, we present a new group of round functions we called the Moirai which allow us to explore what a round function providing optimal performance/security tradeoff could look like.

Self-Sovereign Identity (SSI) frameworks enable individuals to receive and present digital credentials in a user-controlled way. Revocation mechanisms ensure that invalid or withdrawn credentials cannot be misused. These revocation mechanisms must be scalable (e.g., at national scale) and preserve core SSI principles such as privacy, user control, and interoperability. Achieving both is hard, and finding a suitable trade-off remains a key challenge in SSI research. This paper introduces UPPR, a revocation mechanism for One-Show Verifiable Credentials (oVCs) and unlinkable Anonymous Credentials (ACs). Revocations are managed using per-credential Verifiable Random Function (VRF) tokens, which are published in a Bloom filter cascade on a blockchain. Holders prove non-revocation via a VRF proof for oVCs or a single Zero-Knowledge Proof for ACs. The construction prevents revocation status tracking, allows holders to stay offline, and hides issuer revocation behavior. We analyze the privacy properties of UPPR and provide a prototype implementation on Ethereum. Our implementation enables off-chain verification at no cost. On-chain checks cost 0.56-0.84 USD, while issuers pay only 0.00002-0.00005 USD per credential to refresh the revocation state.

ARADI is a low-latency block cipher introduced by the U.S. National Security Agency (NSA) for secure and efficient memory encryption applications. In contrast to most ciphers proposed in the academic community, the design rationale for ARADI has not been publicly disclosed, limiting external evaluation to independent cryptanalysis. Several such analyses have already been published, with the most effective attacks to date reaching up to 12 out of 16 rounds. In this work, we present a differential meet-in-the-middle attack on ARADI that incorporates several new optimizations and dedicated techniques, enabling, for the first time, an attack extending to 14 rounds of the cipher.

Domain-informed deep learning integrates specialized knowledge into deep neural networks through modifications to inputs, architectures, loss functions, or training processes. Such approach is usually employed to improve model performance in low/noisy data regime via domain-specific priors. Side-channel analysis is such a domain and even though straight deep learning has become a powerful tool for profiled side-channel, it still struggles on more complex datasets because of weak signals distributed across many datapoints. Initial attempts at integrating specific knowledge into models have been extremely successful to pushing boundaries of what can be achieved with deep learning. For example, integrating knowledge about the masking scheme in Masure et al provides significant improvements in decreasing the profiling complexity. Marquet et Oswald went further by leveraging redundant features across multiple tasks, such as shared randomness or common masking flow. In this work, we continue the discussion by using belief propagation on a larger graph to guide the learning. We introduce a multi-task learning model that explicitly integrates a factor graph reflecting the algebraic dependencies among intermediates in the computations of Kyber's inverse Number Theoretic Transform (iNTT). Such framework allow the model to learn a joint representation of the related tasks that is mutually beneficial. For the first time, we show that one can perform a belief propagation during training even when one does not have access to the internal randomness, on the masked shares, potentially improving greatly the performances of the attack.

Poseidon and Poseidon2 are cryptographic hash functions crafted for efficient zero-knowledge proof systems and have seen wide adoption in practical applications. We introduce the use of the Graeffe transform in univariate polynomial solving within this line of work. The proposed method streamlines the root recovery process in interpolation attacks and achieves several orders of magnitude acceleration in practical settings, enabling a new and more efficient class of attacks against Poseidon targeting round-reduced permutations and constrained input/output instances. We release open-source code and describe our method in detail, demonstrating substantial improvements over prior approaches: reductions in wall time by a factor of $2^{13}$ and in memory usage by a factor of $2^{4.5}$. Memory-access costs for NTTs turn out to be a dominant barrier in practice. And we prove that this cost increases at least as the $4/3$-power of the input size (up to logarithmic factors), which suggests the commonly used pseudo-linear cost model may underestimate the true resource requirements. This behavior contrasts with multivariate equation solving, whose main bottleneck remains finite-field linear algebra. We argue that, when selecting parameters, designers should account for interpolation-based attacks explicitly, since their practical hardness is determined by different, and sometimes stronger, resource constraints than those of multivariate techniques.

Quantum depth plays a critical role in determining the performance of quantum implementations, yet quantum programming tools often fail to produce depth-optimal compilations of linear layers. In this work, we present a systematic and automated framework that reorders quantum gate sequences of linear layers to obtain depth-efficient quantum implementations. Our method consistently produces circuits with lower depth compared to prior implementations. We apply the framework to a range of cryptographic operations, including the AES MixColumn, internal layers of the AES S-box, binary field squaring, and modular reduction in binary field multiplication. In all these cases, our method achieves meaningful reductions in quantum depth—for example, lowering the depth of the AES MixColumn and S-box circuits. This work explores optimal quantum circuit designs for quantum programming tools, improves the accuracy of quantum resource estimation for cryptanalysis, and supports more realistic evaluations of post-quantum security.

We show that the authentication scheme (IEEE Internet Things J., 24310-24322, 2024) cannot be practically implemented, because it misused the elliptic curve group law---point multiplication, which is represented as $kP$, where $k$ is an integer and $P$ is a point on an elliptic curve. But the scheme uses the false representation $Q_iQ_j$ to construct verification equations, where $Q_i$ and $Q_j$ are two points. Besides, we find that an adversary can retrieve the target relay device's secret key using the intercepted message via open channels.

We construct the first unambiguous succinct non-interactive arguments (SNARGs) for P and incrementally verifiable computation (IVC) for P from the polynomial hardness of learning with errors (LWE). Unambiguity guarantees that it is computationally hard to find two distinct accepting proofs for the same statement. As an application, we establish the first PPAD hardness result based on the polynomial hardness of LWE combined with a widely believed complexity assumption. Central to our approach is a new notion of rate-1 witness-unambiguous batch arguments for NP, which we give the first construction from the polynomial hardness of LWE. This notion may be of independent interest.

Sitting at the core of folding-based recursion, multi-instance accumulation still forces the recursive verifier to combine $\ell$ committed objects, incurring $\Theta(\ell)$ circuit-expensive commitment random linear combinations (CRCs). Here $\ell$ denotes the number of instances accumulated per step, and a CRC is an in-circuit random linear combination of commitments. Eliminating the linear-in-$\ell$ CRC overhead is the central challenge, and we address it with Quasar, reducing in-circuit CRCs from $\Theta(\ell)$ to $O(1)$ per step, while the unavoidable cost of reading and processing public inputs remains linear as usual. Realizing constant-commitment accumulation, Quasar's multi-cast reduction commits to a union polynomial and verifies a random partial evaluation, yielding a reduced instance with $O(1)$ commitments. Moreover, composing multi-cast with a standard $2$-to-$1$ folding reduction yields a multi-instance accumulation scheme whose verifier uses $O(\log \ell)$ field work and $O(1)$ CRCs per step. Across instantiations with suitable polynomial commitments, our instantiation supports linear-time accumulation provers, plausible post-quantum security, and parallelizable proving at each step. Not only does Quasar keep CRCs constant in $\ell$, it also yields a multi-instance incrementally verifiable computation with lower recursion overhead.

Differential meet-in-the-middle attacks, introduced by Boura et al. in 2023, propose a new way of dealing with differential distinguishers. It allows, in particular, to combine differential attacks with initial structures, that were usually used exclusively for meet-in-the-middle attacks. Several applications of this new technique have been published, but so far the results on Feistel constructions have not improved much upon previous best known attacks. In this paper, we apply them on Feistel constructions with all the improvements proposed so far, and we propose some additional new ideas to generically improve these kinds of attacks. We also propose an automatized tool for optimizing the attacks on Simon-like constructions. Our tool outputs a graphical representation of the attack that makes it very easy to verify. All this has allowed us to provide improved single-key key-recovery attacks on most of the variants of Simon, Simeck and CLEFIA-256, that increase the highest number of rounds attacked by 1 or 2 in nearly all the cases.

We study the practicality of a primal hybrid lattice attack on LWE. Despite significant research efforts, the state-of-the-art in practical LWE record computations so far is the plain primal attack with Kannan's embedding. Building on an idea by Espitau and Kirchner, Bernstein proposed in 2023 an LWE hybrid attack that asymptotically outperforms the primal attack. In a nutshell, Bernstein's attack enumerates some coordinates of an LWE key and uses the sophisticated Batch-CVP {\em (Randomized) Slicer} algorithm to solve LWE in a dimension-reduced lattice. The practical implications of this attack however remain widely unclear. One of the major obstacles for judging practicality is the lack of a fast, fully functional Slicer implementation. For the first time, we provide an efficient Slicer implementation that includes all required algorithmic ingredients like locality sensitive hashing. Building on our Slicer implementation, we implement a generalization of Bernstein's algorithm. While Bernstein's attack works only for LWE, ours also applies to a more general BDD setting. Let $(\mathbf{B}, \mathbf{t})$ be a BDD instance, where the target $\mathbf{t}$ is off from the $d$-dimensional lattice $\mathcal{L}(\mathbf{B})$ by some error~$\mathbf{e}$, sampled coordinate-wise from a distribution $\mathcal{D}$. We show that for hard BDD instances, our BDD hybrid asymptotically speeds up primal's complexity of $T=2^{0.292d + o(d)}$ down to $T^{1-\mathcal{K}}$, where $\mathcal{K} \approx \big(1+\frac{H(\mathcal{D})}{0.058}\big)^{-1}$ with $H(\cdot)$ the Shannon entropy. Depending on $\mathcal{D}$, the constant $\mathcal{K}$ can be small, making practical improvements difficult. We test our Slicer-based implementation inside an implementation of our BDD hybrid lattice attack to tackle LWE instances. We choose two ternary LWE secrets with different entropies $H(\mathcal{D})$ as used in NTRU, and the centered binomial distribution as used in Kyber. For all three distributions in all tested LWE dimensions $n \in [160, 210]$, our Slicer-based implementation practically demonstrates measurable speedups over the primal attack, up to a factor of $5$. We also show that for parameters as originally suggested by Regev, the hybrid attack cannot improve over primal.