Abstract In dynamic multi-key Fully Homomorphic Encryption (MFHE) scheme, homomorphic operations on ciphertexts encrypted under different keys are allowed. Moreover, the resulting ciphertext can be further computed with ciphertexts under other keys. In 2016, Peikert and Shiehian presented a dynamic MFHE scheme that relies on using a Common Reference String (CRS) and circular security assumptions. Subsequently, various variants of dynamic MFHE schemes were proposed based on this scheme, but these variant schemes still rely on a CRS and circular security assumption. Therefore, constructing a MFHE GSW scheme relies solely on the LWE assumption and does not require a CRS remains a valuable challenge. This paper introduces a new dynamic MFHE scheme. Compared to existing solutions, our scheme does not require the circular security assumption and depends solely on the LWE assumption, which theoretically provides an enhanced level of security. Additionally, it eliminates the need for a shared CRS but instead replaces the CRS with each party independently generate random matrices. This decentralized approach enhances the ability of users to generate keys independently. The size of ciphertexts and public keys in our scheme is comparable to other existing solutions, thus maintaining efficiency in both computation and ciphertext extension.

Abstract Traditional fully homomorphic encryption(FHE) schemes allow computation only on data encrypted under the same public key. Multi-Key Fully Homomorphic Encryption (MKFHE) enables arbitrary operations on data encrypted with different public keys, allowing all participating users jointly decrypting the final ciphertext. The multi-key BFV FHE scheme inherits BFV’s advantages in ring element encryption and scale invariance. Nonetheless, it also has some disadvantages, such as additional noise generated during the relinearization process, the need for costly transformations during the external product process, and the requirement for a Common Reference String (CRS). In this paper, we investigate the MKFHE scheme for RLWE-based BFV. Firstly, we improve the modulus size of the evaluation key and the public key to construct a modulus enchancement relinearization method, which can significantly reduce the noise generated during the relinearization process. Secondly, we propose to use an inner product via Gadget decomposition in the relinearization based on the MK-BFV scheme instead of the original outer product operation, which can reduce the complexity of the NTT operation to $$\left( {d + 2\tilde{d}} \right)r^{\prime } /\left( {d + 2} \right)\tilde{l}$$ d + 2 d ~ r ′ / d + 2 l ~ of the original one. Finally, we propose a MK-BFV without CRS on the basis of the previous ones, which enhances the user's control over his own key.

With the explosive growth of Internet information, the classification of massive Internet data plays a very important role in real life. Text classification has been widely used in spam text recognition, intention recognition, text matching, named entity recognition, and other fields. At present, many enterprises provide APIs for text classification for users. Users can upload their data to the cloud server deployed by service providers for analysis, and return the final classification results. However, there is a risk of user data and model leakage in this process. To solve this problem, we propose a privacy-preserving text classification scheme using CKKS fully homomorphic encryption scheme and self-attention mechanism model in the multi-party security computing scenario. Our scheme ensures that user can achieve efficient encrypted data analysis under the premise of their data security, and user must be authorized by the service provider to use the model. Finally, compared with the experimental results of the previous research on privacy text classification under fully homomorphic encryption, the implementation improves the accuracy by 7.97% at most and speed-ups 282.4 times for inference at most, and we ensure the security of the protocol participants.

With the exponential expansion of Internet information, technology that combines big data and artificial intelligence has gradually developed. Pre-trained large-scale language models with the transformer architecture as the core have begun to be used in daily life, resulting in the huge market of MLaaS. leading to the significant market of Machine Learning as a Service (MLaaS). Although MLaaS brings huge benefits to users, it requires receiving users’ data for processing, which includes many sensitive data. While MLaaS offers considerable benefits, it necessitates processing users’ data, which includes much sensitive data.Therefore, the problem of privacy data leakage has also been exposed. In this article paper, we propose a novel language model secure inference scheme based on secure multi-party computation (MPC) technology. This solution involves three non-colluding parties: the data provider, the model provider, and the computing power provider. Compared with direct inference on pre-trained large models, the proposed security inference framework improves the inference speed by 1.55-6.25 times. Our findings demonstrate that, when compared to conventional inference methods on pre-trained large-scale models, our approach significantly enhances inference efficiency, achieving speed improvements ranging from 1.55 to 6.25 times.

With the use of a large number of mobile devices, the problem of key leakage becomes more and more serious. In view of the excellent characteristics of lattice cipher and forward-secure digital signature scheme, the construction of identity-based forward-secure digital signature based on lattice technology has become a research hotspot. However, the identity-based forward secure digital signature scheme on the existing grid has the disadvantage of excessive signature length. This paper uses the technique (without trapdoors) of Lyubashevsky and extended Samplepre, an efficient identity-based forward secure signature scheme from lattice is proposed. Its security is based on the Small Integer Solution (SIS) difficulty assumption, and the strong non-forgery of the signature scheme is achieved. Low computing overhead; The analysis results show that, compared with the existing schemes, the key and signature are smaller in size, more efficient in computing, able to resist quantum attacks, and more practical.

We propose a GSW-style fully homomorphic encryption scheme over the integers (FHE-OI) that is more efficient than the prior work by Benarroch et al. (PKC 2017). To reduce the expansion of ciphertexts, our scheme consists of two types of ciphertexts: integers and vectors. Moreover, the computational efficiency in the homomorphic evaluation can be improved by hybrid homomorphic operations between integers and vectors. In particular, when performing vector-integer multiplications, the evaluation has the computational complexity of Ο γ log γ and thus outperforms all prior FHE-OI schemes. To slow down the noise growth in homomorphic multiplications, we introduce a new noise management method called sequentialization; therefore, the noise in the resulting ciphertext increases by a factor of l ⋅ poly λ rather than poly λ l in general multiplications, where l is the number of multiplications. As a result, the circuit with larger multiplicative depth can be evaluated under the same parameter settings. Finally, to further reduce the size of ciphertexts, we apply ciphertext truncation and obtain the integer ciphertext of size Ο λ log λ , thus additionally reducing the size of the vector ciphertext in Benarroch’s scheme from Ο ˜ λ 4 to Ο λ 2 log 2 λ .