ciphertext packing

To address the high online encryption overhead in Fully Homomorphic Encryption (FHE) systems—which critically limits throughput in high-load scenarios such as outsourced databases—this paper proposes a compile-time ciphertext synthesis framework. It shifts ciphertext generation entirely to compilation time via precomputed basis vectors, zero-encryption reuse, and composition of homomorphic addition and scalar multiplication, enabling runtime-zero encryption during data ingestion. We formally define “random-mode homomorphism” for the first time and prove its IND-CPA security via a hybrid game, rigorously characterizing the security boundaries of basis reuse and structured noise injection. The scheme remains compatible with standard FHE APIs while preserving layout semantics for downstream homomorphic operations. Experimental results demonstrate substantial improvements in batch encoding throughput, establishing an efficient, secure, and deployable paradigm for ciphertext injection in high-throughput FHE pipelines.

Homomorphic encryption (HE)-based string matching suffers from high latency and data movement bottlenecks due to ciphertext expansion and expensive homomorphic operations—particularly multiplication and rotation. To address this, we propose a hardware–software co-design: first, a memory-efficient ciphertext packing scheme that eliminates redundant homomorphic operations; second, an in-flash processing (IFP) architecture tailored for HE string matching to enable near-data computing. Compared to state-of-the-art pure-software approaches, our software-only optimization achieves 42.9× speedup and 17.6× energy reduction; integrating IFP further delivers end-to-end acceleration of 136.9× and 256.4× energy savings. This work marks the first application of IFP to HE string matching, establishing a practical and efficient hardware acceleration paradigm for privacy-sensitive applications—including encrypted DNA sequence alignment and secure database search.

To address the low efficiency of plaintext-ciphertext matrix multiplication (PC-MM) under unpacked additive homomorphic encryption (AHE), this work pioneers the adaptation of the Cussen compression-reconstruction algorithm to the unpacked AHE setting, overcoming the performance bottlenecks of conventional schoolbook and Strassen methods. Leveraging an elliptic-curve ElGamal cryptosystem, we integrate SIMD-style data layout optimization with a tree-based reconstruction strategy to enable lightweight deployment on edge devices—specifically, the Raspberry Pi 5. Experiments demonstrate up to a 10× throughput improvement in PC-MM for large matrices and low bit-width operands over state-of-the-art approaches, significantly enhancing the feasibility of privacy-preserving machine learning and encrypted signal processing under resource constraints. Our core contribution is the first compression-reconstruction PC-MM paradigm tailored for unpacked AHE, jointly optimizing security, numerical precision, and edge-device efficiency.

This work addresses the high computational and memory overheads that hinder the practical deployment of fully homomorphic encryption (FHE) in privacy-preserving machine learning, which stem from the inherent complexity of cryptographic operations and inefficient ciphertext packing. To overcome these limitations, the authors propose FEnc², a fragment-based unified encoding framework that treats encrypted tensor layout as a first-class design dimension in FHE systems. By jointly optimizing spatial locality and feature grouping through convolution-aware encoding and architecture-aware ciphertext compression, FEnc² substantially improves slot utilization while reducing rotation complexity and ciphertext count. Built upon the CKKS scheme and compatible with NTT and key-switching accelerations, FEnc² achieves up to 228.83× (GPU) and 226.06× (CPU) speedup for LeNet inference on MNIST, and 4.55× (GPU) and 9.43× (CPU) acceleration for MobileNet inference on ImageNet.

In privacy-sensitive applications such as healthcare, homomorphic encryption (HE)-based private CNN inference suffers from prohibitively high latency in convolutional layers—accounting for over 85% of total inference time. To address this, this paper proposes an efficient HE convolution algorithm. Its core innovations include: (i) the first plaintext multiplication mechanism leveraging Walsh–Hadamard matrices, overcoming the inherent limitation of padding-based convolutions that preclude channel packing in outputs; and (ii) integration of adaptive HE parameter optimization with lazy modular reduction. Experiments on models including VGG-16 demonstrate up to 1.6–3.8× speedup in convolution computation, 2000–8000× compression in encrypted weight storage, 1.3–2.5× reduction in end-to-end inference latency, 2.1–7.9× lower memory footprint, and 1.4–1.5× decrease in communication overhead.

This work addresses the challenge of deploying machine learning in the cloud when legal constraints prohibit sharing sensitive data by proposing a privacy-preserving image classification method based on fully homomorphic encryption (FHE). The authors redesign convolutional neural network architectures to operate efficiently in the encrypted domain, extending the TenSEAL framework to support multi-channel color images, multi-layer convolutions, and average pooling operations, with an efficient inference implementation built upon Microsoft SEAL. Experimental results demonstrate that the proposed approach achieves classification accuracy on MNIST, Kuzushiji-MNIST, Fashion-MNIST, and CIFAR-10 datasets nearly matching that of plaintext models, while maintaining relatively low computational overhead, thereby significantly advancing the practical applicability of FHE to real-world computer vision tasks.

This work addresses the challenge of reducing verification errors and improving computational efficiency in AI decision-making while preserving data privacy. The authors propose a Hash-based Homomorphic Artificial Intelligence (HbHAI) framework that, for the first time, integrates repeated error-correcting codes into homomorphic AI systems and introduces a key-dependent hash function. This design enables native AI algorithms to operate directly on encrypted data, allowing verification error to be reduced arbitrarily and achieving up to 10× data compression. Remarkably, under strong privacy guarantees, the approach reduces both computation time and energy consumption by nearly an order of magnitude—outperforming even plaintext processing in certain scenarios.

This work addresses the high latency introduced by traditional decompression in scientific data analysis, which undermines the storage and transmission benefits of compression. To overcome this limitation, the authors propose a multi-stage, error-bounded decompression and homomorphic analysis framework. By abstracting a generic compression pipeline, the framework enables hierarchical partial decompression and introduces homomorphic operation algorithms tailored to three representative scientific analysis tasks, allowing computations to be performed directly on intermediate compressed representations without full decompression. Implemented atop four mainstream compressors and evaluated across five real-world datasets, the approach consistently reduces data access latency and significantly improves analytical efficiency across diverse workloads.

Fully homomorphic encryption (FHE) remains challenging to execute efficiently due to its substantial computational and memory overhead, exacerbated by a longstanding disconnect between cryptographic optimizations and hardware design. This work proposes a memory-centric, architecture-aware hardware-software co-optimization approach that tightly integrates the CKKS scheme with accelerator design, drastically reducing off-chip memory accesses and temporary data storage. Key innovations include an accelerator-oriented fine-grained coefficient-to-slot transformation, plaintext compression, intermediate modulus upping, dedicated on-chip buffers, and extended functional units. These synergistic enhancements enable, for the first time, sub-millisecond CKKS bootstrapping. Compared to the state-of-the-art FHE accelerators, the proposed design achieves 1.38× to 8.74× higher performance per unit area.

This work addresses the high computational cost of sparse matrix-vector multiplication under homomorphic encryption, which stems from irregular distributions of non-zero elements leading to numerous cyclic diagonals. The authors propose the two-dimensional diagonal packing problem (2DPP) together with a homomorphic encryption-aware cost model, aiming to minimize the number of cyclic diagonals through row and column permutations. Their approach integrates graph bandwidth reduction, anti-bandwidth maximization, spectral ordering, and iterative 2OPT/3OPT refinements, augmented by a dense row/column elimination mechanism to isolate highly dense substructures. Evaluated on 175 SuiteSparse matrices, the method reduces the number of diagonals by an average of 5.5× (up to 45.6×), and when combined with dense elimination, achieves up to a 23.7× reduction in encrypted multiplication cost.