mlsanityData to 7 Oct 2026

HEMET: A Homomorphic-Encryption-Friendly Privacy-Preserving Mobile Neural Network Architecture

arXiv 2106.00038 ↗First posted 31 May 2021cs.CR cs.AIcited by 24

Not on the map. Why?

Abstract (arXiv)

Recently Homomorphic Encryption (HE) is used to implement Privacy-Preserving Neural Networks (PPNNs) that perform inferences directly on encrypted data without decryption. Prior PPNNs adopt mobile network architectures such as SqueezeNet for smaller computing overhead, but we find na"ively using mobile network architectures for a PPNN does not necessarily achieve shorter inference latency. Despite having less parameters, a mobile network architecture typically introduces more layers and increases the HE multiplicative depth of a PPNN, thereby prolonging its inference latency. In this paper, we propose a HE-friendly privacy-preserving Mobile neural nETwork architecture, HEMET. Experimental results show that, compared to state-of-the-art (SOTA) PPNNs, HEMET reduces the inference latency by 59.3%~ 61.2%, and improves the inference accuracy by 0.4 % ~ 0.5%.

Key facts. Cited by 24 papers in the map (data through 7 Oct 2026). Builds on 7 of its 13 references that are in the map. First posted on arXiv 31 May 2021.

Builds on

7 of 13 references.

References · 7

  1. 2015Batch Normalization: Accelerating Deep Network Training by Reducing Internal Covariate Shiftcited by 6,979
  2. 2016Inception-v4, Inception-ResNet and the Impact of Residual Connections on Learningcited by 1,509
  3. 2016SqueezeNet: AlexNet-level accuracy with 50x fewer parameters and <0.5MB model sizecited by 1,054
  4. 2018Gazelle: A Low Latency Framework for Secure Neural Network Inferencecited by 193
  5. 2019EVA: An Encrypted Vector Arithmetic Language and Compiler for Efficient Homomorphic Computationcited by 20
  6. 2020NASS: Optimizing Secure Inference via Neural Architecture Searchcited by 7
  7. 2020AutoPrivacy: Automated Layer-wise Parameter Selection for Secure Neural Network Inferencecited by 7

Built on by

Cited by year (2026 to 7 Oct)

Citing papers per year: 2022: 4; 2023: 4; 2024: 5; 2025: 6; 2026: 5.

24 papers.

Sort:

By area:

  1. 2022Scaling up Trustless DNN Inference with Zero-Knowledge Proofscited by 30
  2. 2023AutoFHE: Automated Adaption of CNNs for Efficient Evaluation over FHEcited by 15
  3. 2022AESPA: Accuracy Preserving Low-degree Polynomial Activation for Fast Private Inferencecited by 14
  4. 2024Verifiable evaluations of machine learning models using zkSNARKscited by 14
  5. 2024Trustless Audits without Revealing Data or Modelscited by 14
  6. 2023HyPHEN: A Hybrid Packing Method and Optimizations for Homomorphic Encryption-Based Neural Networkscited by 9
  7. 2025TeleSparse: Practical Privacy-Preserving Verification of Deep Neural Networkscited by 9
  8. 2022Private and Reliable Neural Network Inferencecited by 6
  9. 2023Training Large Scale Polynomial CNNs for E2E Inference over Homomorphic Encryptioncited by 5
  10. 2023Falcon: Accelerating Homomorphically Encrypted Convolutions for Efficient Private Mobile Network Inferencecited by 5
  11. 2024ObfuscaTune: Obfuscated Offsite Fine-tuning and Inference of Proprietary LLMs on Private Datasetscited by 4
  12. 2025TFHE-Coder: Evaluating LLM-agentic Fully Homomorphic Encryption Code Generationcited by 4
12 more
  1. 2024CryptoTrain: Fast Secure Training on Encrypted Datasetcited by 3
  2. 2022Efficient Pruning for Machine Learning Under Homomorphic Encryptioncited by 2
  3. 2025CryptoFace: End-to-End Encrypted Face Recognitioncited by 2
  4. 2024Self-Supervised Inference of Agents in Trustless Environmentscited by 1
  5. 2025Privacy-Preserving Chest X-ray Classification in Latent Space with Homomorphically Encrypted Neural Inferencecited by 1
  6. 2025DictPFL: Efficient and Private Federated Learning on Encrypted Gradientscited by 1
  7. 2026CGF-Softmax: A Cumulant-Based Softmax Reformulation for Efficient Inference under Homomorphic Encryptioncited by 1
  8. 2025FastFHE: Packing-Scalable and Depthwise-Separable CNN Inference Over FHEcited by 0
  9. 2026RobPI: Robust Private Inference against Malicious Clientcited by 0
  10. 2026CausShield: Sample Reconstruction-Resilient Vertical FL via Causal Representation Learningcited by 0
  11. 2026FEnc2: Unifying Data Packing for Efficient Private Inference via Convolution and Architecture-Aware Fragment Encodingcited by 0
  12. 2026Repurposing CLIP to Localize at Pixel Levelcited by 0