~ similar to 2610.26442· 20 results
The paper proves that platform-deterministic inference is a necessary and sufficient condition for trustworthy AI, establishing that AI trust fundamentally relies on consistent arithmetic.
Badih Ghazi, Cristóbal Guzmán, Pritish Kamath, Alexander Knop +2 more
The paper studies the problem of generating synthetic data under differential privacy and achieves fixed-parameter tractability.
This paper studies the benefits and limitations of feedback in computation, specifically in the context of deterministic programs with random clock movements.
This paper provides a comprehensive, system-level comparison of MPC and FHE for Privacy-Preserving Machine Learning (PPML) across various models and environments, moving beyond single-metric latency a…
The paper introduces PolyVeil, a protocol for private Boolean summation that uses permutation matrices in the Birkhoff polytope, achieving strong security guarantees while highlighting a fundamental t…
The paper studies the resources required to batch verify Boolean functions and provides lower bounds on the witness-query tradeoff based on approximate degree.
The paper provides the first machine-checked universal proof, using ring theory, that value-independence implies identical marginal distributions for arithmetic masking, thereby extending the verifica…
The paper proposes a provably secure, single-round two-party computation protocol for approximate matrix multiplication using lattice-based cryptography, demonstrated for secure control law implementa…
StateSync-GKR is presented, a reusable proof engine that conserves memory and reduces repeated setup for authenticated state transitions.
This paper introduces IncrementalExecution, a novel online framework for optimising the number of shots needed to meet a target accuracy in quantum algorithms.
The paper quantifies the cost of privacy in language identification and generation using differentially private (DP) methods, finding that the cost is surprisingly mild, particularly absent under appr…
TAPAS introduces an efficient, asymmetric two-server private aggregation scheme that significantly reduces computational and communication costs for large-scale federated learning compared to existing…
The paper proves that for resources with structural parallelizability (like divisibility and transferability), it is impossible to enforce a linear cost for concentrating influence, demonstrating that…
The paper introduces the PML envelope, a novel definition that provides a robust and operationally meaningful measure of information leakage about a secret, satisfying both post-processing robustness…
This paper proposes Interflow, an algorithm-hardware co-design that improves the efficiency of privacy-preserving machine learning by pipelining correlated randomness generation with online execution.
The paper introduces a lightweight, sampling-based cryptographic protocol for verifiable AI inference that drastically reduces proving overhead from minutes to milliseconds by leveraging statistical p…
The paper develops a quantum analogue of interactive oracle proofs (qIOP) for any language in QMA, with polynomial communication and polylogarithmic qubit queries.
The paper presents a new method for privately releasing a k-dimensional vector under updates, improving the time complexity of sampling a noise vector from the binary tree mechanism with Gaussian nois…