mlsanityData to 14 Oct 2026

Nimbus: Secure and Efficient Two-Party Inference for Transformers

arXiv 2411.15707 ↗First posted 24 Nov 2024cs.CR cs.AIcited by 16

Not on the map. Why?

Abstract (arXiv)

Transformer models have gained significant attention due to their power in machine learning tasks. Their extensive deployment has raised concerns about the potential leakage of sensitive information during inference. However, when being applied to Transformers, existing approaches based on secure two-party computation (2PC) bring about efficiency limitations in two folds: (1) resource-intensive matrix multiplications in linear layers, and (2) complex non-linear activation functions like mathsfGELU and mathsfSoftmax. This work presents a new two-party inference framework mathsfNimbus for Transformer models. For the linear layer, we propose a new 2PC paradigm along with an encoding approach to securely compute matrix multiplications based on an outer-product insight, which achieves 2.9× ~ 12.5× performance improvements compared to the state-of-the-art (SOTA) protocol. For the non-linear layer, through a new observation of utilizing the input distribution, we propose an approach of low-degree polynomial approximation for mathsfGELU and mathsfSoftmax, which improves the performance of the SOTA polynomial approximation by 2.9× ~ 4.0×, where the average accuracy loss of our approach is 0.08% compared to the non-2PC inference without privacy. Compared with the SOTA two-party inference, mathsfNimbus improves the end-to-end performance of bert inference by 2.7× ~ 4.7× across different network settings.

Key facts. Cited by 16 papers in the map (data through 14 Oct 2026). Builds on 24 of its 46 references that are in the map. First posted on arXiv 24 Nov 2024.

Builds on

24 of 46 references.

References · 24

  1. 2017Attention Is All You Needcited by 60,361
  2. 2018BERT: Pre-training of Deep Bidirectional Transformers for Language Understandingcited by 40,399
  3. 2020An Image is Worth 16x16 Words: Transformers for Image Recognition at Scalecited by 30,186
  4. 2023LLaMA: Open and Efficient Foundation Language Modelscited by 14,132
  5. 2018GLUE: A Multi-Task Benchmark and Analysis Platform for Natural Language Understandingcited by 4,899
  6. 2016Gaussian Error Linear Units (GELUs)cited by 3,502
  7. 2019HuggingFace's Transformers: State-of-the-art Natural Language Processingcited by 2,218
  8. 2022SmoothQuant: Accurate and Efficient Post-Training Quantization for Large Language Modelscited by 1,386
  9. 2023Deja Vu: Contextual Sparsity for Efficient LLMs at Inference Timecited by 252
  10. 2018Gazelle: A Low Latency Framework for Secure Neural Network Inferencecited by 193
  11. 2021CrypTen: Secure Multi-Party Computation Meets Machine Learningcited by 131
  12. 2023OliVe: Accelerating Large Language Models via Hardware-friendly Outlier-Victim Pair Quantizationcited by 92
12 more
  1. 2020CrypTFlow2: Practical 2-Party Secure Inferencecited by 82
  2. 2022MPCFormer: fast, performant and private Transformer inference with MPCcited by 69
  3. 2018Low Latency Privacy Preserving Inferencecited by 57
  4. 2022THE-X: Privacy-Preserving Transformer Inference with Homomorphic Encryptioncited by 57
  5. 2022ANT: Exploiting Adaptive Numerical Data Type for Low-bit Deep Neural Network Quantizationcited by 51
  6. 2023ChatGPT: A Meta-Analysis after 2.5 Monthscited by 51
  7. 2023PUMA: Secure Inference of LLaMA-7B in Five Minutescited by 48
  8. 2021SIRNN: A Math Library for Secure RNN Inferencecited by 28
  9. 2022Transkimmer: Transformer Learns to Layer-wise Skimcited by 24
  10. 2022MPCViT: Searching for Accurate and Efficient MPC-Friendly Vision Transformer with Heterogeneous Attentioncited by 20
  11. 2021Block-Skim: Efficient Question Answering for Transformercited by 18
  12. 2024Ditto: Quantization-aware Secure Inference of Transformers upon MPCcited by 8

Built on by

Cited by year (2026 to 14 Oct)

Citing papers per year: 2024: 1; 2025: 7; 2026: 8.

16 papers.

Sort:

By area:

  1. 2024CENTAUR: Bridging the Impossible Trinity of Privacy, Efficiency, and Performance in Privacy-Preserving Transformer Inferencecited by 6
  2. 2025Cape: Context-Aware Prompt Perturbation Mechanism with Differential Privacycited by 5
  3. 2025An Efficient Private GPT Never Autoregressively Decodescited by 2
  4. 2025An Attack to Break Permutation-Based Private Third-Party Inference Schemes for LLMscited by 2
  5. 2025ENSI: Efficient Non-Interactive Secure Inference for Large Language Modelscited by 2
  6. 2026SecMoE: Communication-Efficient Secure MoE Inference via Select-Then-Computecited by 2
  7. 2026On the (In-)Security of the Shuffling Defense in the Transformer Secure Inferencecited by 2
  8. 2025Cascade: Token-Sharded Private LLM Inferencecited by 1
  9. 2025Black Box Absorption: LLMs Undermining Innovative Ideascited by 1
  10. 2026OSNIP: Breaking the Privacy-Utility-Efficiency Trilemma in LLM Inference via Obfuscated Semantic Null Spacecited by 1
  11. 2026MOSAIC: Masked Outsourcing of Secure AI Computationscited by 1
  12. 2025PubSub-VFL: Towards Efficient Two-Party Split Learning in Heterogeneous Environments via Publisher/Subscriber Architecturecited by 0
4 more
  1. 2026Privacy-Preserving Mechanisms Enable Cheap Verifiable Inference of LLMscited by 0
  2. 2026A Novel Semantic Manifold Alignment Attack against Embedding-to-Embedding Obfuscation in Privacy-Preserving LLMscited by 0
  3. 2026OptiPrime: Optimizing Private Inference through Protocol-Hardware Co-designcited by 0
  4. 2026A Graph-Based Analysis of Comparability, Scale, and Attribution in Private Transformer Inferencecited by 0