Generalized BGV, BFV, and CKKS for Homomorphic Encryption over Matrix Rings¶
Authors: Mali
Year: 2025
Venue: IACR ePrint 2025/972
type: scheme collection: schemes status: auto
Links: ePrint 2025/972
Visual lesson: Generalized BGV/BFV/CKKS over matrix rings · Scheme table row: Generalized BGV/BFV/CKKS over matrix rings
Summary¶
Auto-extracted summary (unreviewed)
This ePrint generalises the BGV-like schemes (BGV, BFV, CKKS) so that plaintext and ciphertext spaces are matrix rings of arbitrary dimension, giving native encrypted matrix multiplication without the slot rotations required by SIMD packing approaches. A new superoperator technique represents linear and quadratic expressions in the secret key despite noncommutativity, enabling relinearization after multiplication; security rests on Module-LWE with module rank equal to the matrix dimension. The author argues that Ring-LWE, Module-LWE and LWE are potentially equally efficient for HE in information density and noise growth, differing only in matrix size.
Abstract¶
Some of the most valuable applications of homomorphic encryption, such as encrypted machine learning inference, require efficient large-scale plaintext-ciphertext and ciphertext-ciphertext matrix multiplications. Current state-of-the-art techniques for matrix multiplications all build on the ability to pack many ciphertexts into a ciphertext and compute on them in a Single Instruction, Multiple Data (SIMD) manner. However, to fit the operation of matrix multiplication into this computational model, a large number of additional costly operations need to be performed, such as the rotation of elements between the plaintext slots. In this work, we propose an orthogonal approach to performing encrypted matrix operations with BGV-like encryption schemes, where the plaintext and ciphertext spaces are generalized to a matrix ring of arbitrary dimension. To deal with the inherent problem of noncommutativity in the case of matrix rings, we present a new superoperator technique to better represent linear and quadratic expressions in the secret key, which allows for the relinearization of ciphertexts after multiplication. The security of the modified encryption schemes is based on Module-LWE with module rank equal to the dimension of the matrices. With this construction, we demonstrate that Ring-LWE, Module-LWE, and LWE are potentially equally efficient for homomorphic encryption, both in terms of useful information density and noise growth, only for different sizes of matrices. 1 Contents
Source: pdf
Extracted details¶
| Field | Value | Source | Confidence |
|---|---|---|---|
| contribution | matrix-ring generalisation of BGV, BFV and CKKS | llm | 0.70 |
| contribution | superoperator technique for relinearization over noncommutative matrix rings | llm | 0.70 |
| contribution | compact homomorphic encryption scheme based on Module-LWE | llm | 0.70 |
| contribution | analysis of ring expansion factor and noise growth | llm | 0.70 |
| contribution | argument that LWE, Module-LWE and Ring-LWE are equally efficient for HE at different matrix sizes | llm | 0.70 |
| schemes | BGV | llm | 0.70 |
| schemes | BFV | llm | 0.70 |
| schemes | CKKS | llm | 0.70 |
| schemes | Module-LWE | llm | 0.70 |
| tasks | encrypted matrix multiplication | llm | 0.70 |
| tasks | encrypted machine learning inference (motivating application) | llm | 0.70 |
| interaction | non-interactive FHE | llm | 0.70 |
| limitations | third-generation schemes such as TFHE/CGGI are not covered | llm | 0.70 |
| llm relevance | low | llm | 0.70 |
Scheme card: Generalized BGV/BFV/CKKS over matrix rings¶
| Group | 4. Matrix-native FHE |
| Kind | Scheme |
| Hardness assumption | Matrix-ring LWE |
| Plaintext type | Exact or approximate matrices |
| SIMD packing | Native matrices |
| Bootstrapping | Yes |
| Post-quantum | Yes |
| Status | Active SOTA (emerging) |
| Fit for LLM inference | Native plaintext-ciphertext and ciphertext-ciphertext matrix products without slot packing. |
Keyword mentions¶
Counts of vocabulary terms in the full text (first page of occurrence in brackets).
- scheme: BGV 24 [p.1], CKKS 22 [p.1], BFV 19 [p.1], GSW 4 [p.4], Leveled / somewhat HE 4 [p.12], TFHE 3 [p.3], NTRU 2 [p.20]
- library: HElib 3 [p.4]
- technique: Relinearization 32 [p.1], Matrix multiplication 20 [p.1], Bootstrapping 7 [p.3], SIMD packing 4 [p.1], Rotations 3 [p.1], Key switching 2 [p.5]
- assumption: LWE 98 [p.1], RLWE 35 [p.1], Lattice-based 15 [p.3]
- task: Language modeling / generation 4 [p.6]
- topic: Homomorphic encryption 21 [p.1]
Files¶
| Path (under papers/) | Collection | Pages | Size | SHA-256 | Note |
|---|---|---|---|---|---|
references/schemes/s27-0_generalized-bgv-bfv-and-ckks-for-homomorphic-encryption-over.pdf |
schemes | 20 | 402.4 KB | 1726b619bef0 |
Database key: generalized-bgv-bfv-and-ckks-for-homomorphic-encryption-over-matrix · created 2026-11-05 · updated 2026-11-05 · edit with python3 tools/paperdb.py set generalized-bgv-bfv-and-ckks-for-homomorphic-encryption-over-matrix COLUMN VALUE