Abstract Over the past few decades, we have seen a proliferation of advanced cryptographic primitives with lossy or homomorphic properties built from various assumptions such as Quadratic Residuosity, Decisional Diffie–Hellman, and Learning with Errors. These primitives imply hard problems in the complexity class $$\mathcal {SZK}$$ SZK (statistical zero-knowledge); as a consequence, they can only be based on assumptions that are broken in $$\mathcal {BPP}^{\mathcal {SZK}}$$ BPP SZK . This poses a barrier for building advanced cryptography from code-based assumptions such as Learning Parity with Noise (LPN), as LPN is only known to be in $$\mathcal {BPP}^{\mathcal {SZK}}$$ BPP SZK under an extremely low noise rate $$\frac{\log ^2 n}{n}$$ log 2 n n , for which it is broken in quasi-polynomial time. In this work, we propose a new code-based assumption: Dense-Sparse LPN, that falls in the complexity class $$\mathcal {BPP}^{\mathcal {SZK}}$$ BPP SZK and we conjecture to be secure against subexponential time adversaries. Our assumption is a variant of LPN that is inspired by McEliece’s cryptosystem and the random $$k\text{- }$$ k - XOR problem in average-case complexity. Roughly, the assumption states that $$\begin{aligned}({\textbf{T}}\, {\textbf{M}}, {\textbf{s}} \,{\textbf{T}}\, {\textbf{M}} + {\textbf{e}}) \quad \text {is indistinguishable from}\quad ({\textbf{T}} \,{\textbf{M}}, {\textbf{u}}),\end{aligned}$$ ( T M , s T M + e ) is indistinguishable from ( T M , u ) , for a random (dense) matrix $${\textbf{T}}$$ T , random sparse matrix $${\textbf{M}}$$ M , and sparse noise vector $${\textbf{e}}$$ e drawn from the Bernoulli distribution with inverse polynomial noise rate. We leverage our assumption to build lossy trapdoor functions (Peikert-Waters STOC 08). This gives the first post-quantum alternative to the lattice-based construction in the original paper. Lossy trapdoor functions, being a fundamental cryptographic tool, are known to enable a broad spectrum of both lossy and non-lossy cryptographic primitives; our construction thus implies these primitives in a generic manner. In particular, we achieve collision-resistant hash functions with plausible subexponential security, improving over a prior construction from LPN with noise rate $$\frac{\log ^2 n}{n}$$ log 2 n n

Parkinson's disease (PD) is a progressive neurodegenerative disorder affecting millions worldwide, characterized by a wide range of motor and non-motor symptoms. Among these symptoms, alterations in speech and voice quality stand out as early and prominent indicators of the disease. Recently, the emergence of speech foundation models has revolutionized the field by providing powerful tools for speech processing and feature extraction. In this article, we investigate the capabilities of three state-of the art speech foundation models, wav2vec2.0, Whisper and SeamlessM4T, to develop robust and accurate methods for PD detection from voice recordings. We experiment with both direct feature extraction and finetuning of the foundation models for the PD classification task, and validate the results against clinical and neuroimaging data. We achieve promising results using both pretrained features and models' finetuning, with finetuning providing stronger performance, up to 91.35% for AUC, which is the new state of the art on the ICEBERG dataset. The predictions of our models also show good correlation with clinical as well as DaTSCAN scores, proving the feasibility to apply speech foundation models for detection of early PD.

A flurry of excitement amongst researchers and practitioners has produced modern proof systems built using novel technical ideas and seeing rapid deployment, especially in cryptocurrencies. Most of these modern proof systems use the Fiat-Shamir (F-S) transformation, a seminal method of removing interaction from a protocol with a public-coin verifier. Some prior work has shown that incorrectly applying F-S (i.e., using the so-called "weak" F-S transformation) can lead to breaks of classic protocols like Schnorr’s discrete log proof; however, little is known about the risks of applying F-S incorrectly for modern proof systems seeing deployment today.In this paper, we fill this knowledge gap via a broad theoretical and practical study of F-S in implementations of modern proof systems. We perform a survey of open-source implementations and find 30 weak F-S implementations affecting 12 different proof systems. For four of these—Bulletproofs, Plonk, Spartan, and Wesolowski’s VDF—we develop novel knowledge soundness attacks accompanied by rigorous proofs of their efficacy. We perform case studies of applications that use vulnerable implementations, and demonstrate that a weak F-S vulnerability could have led to the creation of unlimited currency in a private smart contract platform. Finally, we discuss possible mitigations and takeaways for academics and practitioners.