arXiv subjects

Hyung Tae Lee

Publications and source records attributed to Hyung Tae Lee.

3 recordsLinked to original sources

arXiv2026

Shuffling is Not Enough: Breaking Permutation-Based Model Confidentiality in Hybrid FHE Inference

Hybrid fully homomorphic encryption~(FHE) inference improves the practicality of private inference by letting the server evaluate linear layers homomorphically while the client decrypts and applies nonlinearities. Recent schemes attempt to protect model confidentiality by returning noisy, output-permuted responses and appealing to shuffle-model differential privacy~(DP). We show that this protection fails in the correctness regime required by hybrid FHE systems. For a $d$-input linear layer, $d+1$ admissible queries suffice for exact recovery of a permutation-invariant layer summary, hence for perfect model distinguishability. We further show that input DP is orthogonal to model confidentiality and that the local-DP premise required for shuffle amplification cannot hold under correctness-bounded noise. We recover all linear layers of a \safhire{}-style ResNet-20 end-to-end from TFHE transcripts with zero error, using $d+1$ queries per layer for a total of $5{,}712$ direct queries. Under the same query model, we also confirm exact per-layer recovery on pretrained ImageNet-scale CNNs and ViT-B/16. The leaked spectra enable fingerprinting, lineage attribution, and improved logit-based extraction, while suppressing them destroys inference utility.

cs.CR↗

arXiv2021

IronMask: Modular Architecture for Protecting Deep Face Template

Convolutional neural networks have made remarkable progress in the face recognition field. The more the technology of face recognition advances, the greater discriminative features into a face template. However, this increases the threat to user privacy in case the template is exposed. In this paper, we present a modular architecture for face template protection, called IronMask, that can be combined with any face recognition system using angular distance metric. We circumvent the need for binarization, which is the main cause of performance degradation in most existing face template protections, by proposing a new real-valued error-correcting-code that is compatible with real-valued templates and can therefore, minimize performance degradation. We evaluate the efficacy of IronMask by extensive experiments on two face recognitions, ArcFace and CosFace with three datasets, CMU-Multi-PIE, FEI, and Color-FERET. According to our experimental results, IronMask achieves a true accept rate (TAR) of 99.79% at a false accept rate (FAR) of 0.0005% when combined with ArcFace, and 95.78% TAR at 0% FAR with CosFace, while providing at least 115-bit security against known attacks.

cs.CV↗

arXiv2020

Provably Secure Group Signature Schemes from Code-Based Assumptions

We solve an open question in code-based cryptography by introducing two provably secure group signature schemes from code-based assumptions. Our basic scheme satisfies the CPA-anonymity and traceability requirements in the random oracle model, assuming the hardness of the McEliece problem, the Learning Parity with Noise problem, and a variant of the Syndrome Decoding problem. The construction produces smaller key and signature sizes than the previous group signature schemes from lattices, as long as the cardinality of the underlying group does not exceed $2^{24}$, which is roughly comparable to the current population of the Netherlands. We develop the basic scheme further to achieve the strongest anonymity notion, i.e., CCA-anonymity, with a small overhead in terms of efficiency. The feasibility of two proposed schemes is supported by implementation results. Our two schemes are the first in their respective classes of provably secure groups signature schemes. Additionally, the techniques introduced in this work might be of independent interest. These are a new verifiable encryption protocol for the randomized McEliece encryption and a novel approach to design formal security reductions from the Syndrome Decoding problem.

cs.CR↗