.jpg)
Non-Interactive Zero-Knowledge Arguments with Certified Deletion by Kasra Abbaszadeh of @UofMaryland, and @jon_katz of @Google: https://eprint.iacr.org/2024/1848
This paper introduces non-interactive zero-knowledge arguments with certified deletion, a primitive that lets recipients verify a proof and later provably delete it using a classical certificate. The authors formalize the model, define security guarantees, and present constructions under standard assumptions. One design requires quantum provers and verifiers, while an LWE-based extension allows a classical prover. The work also applies the concept to signatures of knowledge and anonymous credentials with revocation via deletion rather than expiration or blocklists.
Zero-Knowledge Protocols with PVC Security: Striking the Balance between Security and Efficiency by Yi Liu, Yipeng Song, Anjia Yang, and Junzuo Lai of @JinanUniversity: https://eprint.iacr.org/2025/2146
This paper presents a new approach to ZK protocols by introducing publicly verifiable covert (PVC) security as a way to balance security and efficiency. It formally defined ZK with PVC guarantees and proposed a generic transformation for Sigma protocols with 1-bit challenges. The work showed that, for reputation-conscious participants, high cheating deterrence can be achieved with far fewer protocol repetitions, significantly reducing computational and communication overhead in practical settings.
A New Approach to Arguments of Quantum Knowledge by Ruta Jawale, and Kabir Tomer of @HFES_Illini, James Bartusek of @Columbia, and Justin Raizes of @NttResearch: https://eprint.iacr.org/2025/2155
This paper reveals a new framework for publicly verifiable non-interactive zero-knowledge arguments for QMA with a transparent setup based only on a uniformly random string. The authors introduce the first argument of quantum knowledge in this setting, enabling direct extraction of valid quantum witnesses from accepting proofs. The construction combines a strengthened ZX-based QMA verifier, coset-state authentication, and a refined heuristic use of post-quantum obfuscation, with extensions to the quantum pseudorandom oracle model.
Lattice-Based Linkable Ring Signatures for Anonymous and Accountable Whistleblowing by Vishal Pareek, Aditi Kar Gangopadhyay, and Sugata Gangopadhyay of @iitroorkee: https://eprint.iacr.org/2025/2170
This paper describes a lattice-based linkable ring signature scheme designed for anonymous yet accountable whistleblowing in post-quantum settings. It combines exact zero-knowledge proofs, a weak pseudorandom function, and a Merkle tree accumulator to enable signer anonymity while allowing repeated actions to be linked. The construction avoids trapdoors, analyzes concrete signature sizes, and demonstrates how organizations can prevent abuse without revealing identities.
Turning Simulation into Construction: New Uses of NIZK Simulators by @stephan_krenn of @AITtomorrow2day, Kai Samelin of Independent, and Daniel Slamanig of @unibw_m: https://eprint.iacr.org/2025/2194
This work highlights a shift in how non-interactive zero-knowledge simulators can be used, treating simulation not just as a proof technique but as a constructive design tool. The paper shows how simulated proofs of false statements can underpin real cryptographic primitives, including unforgeable signatures, designated verifier signatures, and chameleon hash functions. By embedding simulation directly into protocols, the authors achieve strong security and privacy guarantees under standard assumptions, without relying on heavy or trusted setups.
Consistency Verification for Zero-Knowledge Virtual Machine on Circuit-Irrelevant Representation by Jingyu Ke and Guoqiang Li of @sjtu1896, and Boxuan Liang of Taiyuan University of Technology: https://eprint.iacr.org/2025/2204
This paper discusses a formal approach to verifying consistency in zkVMs that rely on tabular constraint systems rather than circuit-level encodings. It introduces a circuit-irrelevant semantic model and an inductive, row-wise verification method that checks alignment between witness generation and constraints without full trace expansion. The approach is implemented in ZIVER and evaluated on SP1 by @SuccinctLabs components, showing scalable detection of subtle consistency issues in real zkVM designs.
Architecture-private Zero-knowledge Proof of Neural Networks by Yanpei Guo, @qu_wenjie, and @jiahengzhang96 of @NUSingapore, and Zhanpeng Guo of @XDUofChina: https://eprint.iacr.org/2025/2211
In this paper, the authors introduce an architecture-private zero-knowledge proof framework for neural networks, addressing a key gap in zkML where model architectures are typically exposed. They propose a parametrized R1CS and a proof of functional relation to hide CNN architectures while preserving verifiability. Experimental results on VGG models show only modest prover overhead compared to state-of-the-art public-architecture schemes, demonstrating practical feasibility of architecture-level privacy in zkML.
