# HETAL: Efficient Privacy-preserving Transfer Learning with Homomorphic Encryption

arXiv: https://arxiv.org/abs/2403.14111 · HTML: https://mlsanity.com/paper/2403.14111/

## Key facts

- Cited by 12 papers in the map (data through 14 Oct 2026).
- Builds on 14 of its 57 references that are in the map.
- First posted on arXiv 21 Mar 2024.

## Builds on (most cited first, 14 of 14)

- [An Image is Worth 16x16 Words: Transformers for Image Recognition at Scale](https://mlsanity.com/paper/2010.11929/) (2020)
- [MPNet: Masked and Permuted Pre-training for Language Understanding](https://mlsanity.com/paper/2004.09297/) (2020)
- [MedMNIST v2 -- A large-scale lightweight benchmark for 2D and 3D biomedical image classification](https://mlsanity.com/paper/2110.14795/) (2021)
- [Snips Voice Platform: an embedded Spoken Language Understanding system for private-by-design voice interfaces](https://mlsanity.com/paper/1805.10190/) (2018)
- [Information Leakage in Embedding Models](https://mlsanity.com/paper/2004.00053/) (2020)
- [On the Reconstruction of Face Images from Deep Face Templates](https://mlsanity.com/paper/1703.00832/) (2017)
- [Privacy-Preserving Machine Learning with Fully Homomorphic Encryption for Deep Neural Network](https://mlsanity.com/paper/2106.07229/) (2021)
- [More Than Privacy: Applying Differential Privacy in Key Areas of Artificial Intelligence](https://mlsanity.com/paper/2008.01916/) (2020)
- [One-vs-Each Approximation to Softmax for Scalable Estimation of Probabilities](https://mlsanity.com/paper/1609.07410/) (2016)
- [PrivFT: Private and Fast Text Classification with Homomorphic Encryption](https://mlsanity.com/paper/1908.06972/) (2019)
- [Integrating Remote Attestation with Transport Layer Security](https://mlsanity.com/paper/1801.05863/) (2018)
- [SEALion: a Framework for Neural Network Inference on Encrypted Data](https://mlsanity.com/paper/1904.12840/) (2019)
- [Privacy-Preserving Text Classification on BERT Embeddings with Homomorphic Encryption](https://mlsanity.com/paper/2210.02574/) (2022)
- [CryptoTL: Private, Efficient and Secure Transfer Learning](https://mlsanity.com/paper/2205.11935/) (2022)

## Built on by (most cited first, 12 of 12)

- [LLM Post-Training: A Deep Dive into Reasoning Large Language Models](https://mlsanity.com/paper/2502.21321/) (2025)
- [Encryption-Friendly LLM Architecture](https://mlsanity.com/paper/2410.02486/) (2024)
- [CryptoTrain: Fast Secure Training on Encrypted Dataset](https://mlsanity.com/paper/2409.16675/) (2024)
- [AEGIS: Scaling Long-Sequence Homomorphic Encrypted Transformer Inference via Hybrid Parallelism on Multi-GPU Systems](https://mlsanity.com/paper/2604.03425/) (2026)
- [LDPKiT: Superimposing Remote Queries for Privacy-Preserving Distillation](https://mlsanity.com/paper/2405.16361/) (2024)
- [ReBoot: Encrypted Training of Deep Neural Networks with CKKS Bootstrapping](https://mlsanity.com/paper/2506.19693/) (2025)
- [Traveling Salesman-Based Token Ordering Improves Stability in Homomorphically Encrypted Language Models](https://mlsanity.com/paper/2510.12343/) (2025)
- [CGF-Softmax: A Cumulant-Based Softmax Reformulation for Efficient Inference under Homomorphic Encryption](https://mlsanity.com/paper/2602.01621/) (2026)
- [Homomorphic WiSARDs: Efficient Weightless Neural Network training over encrypted data](https://mlsanity.com/paper/2403.20190/) (2024)
- [Efficient Privacy-Preserving KAN Inference Using Homomorphic Encryption](https://mlsanity.com/paper/2409.07751/) (2024)
- [Securing Transfer-Learned Networks with Reverse Homomorphic Encryption](https://mlsanity.com/paper/2505.14323/) (2025)
- [HEAT: Faster Fully Homomorphic Inference via Approximations-Weights Co-Adaptation](https://mlsanity.com/paper/2609.01730/) (2026)

## Abstract (arXiv)

Transfer learning is a de facto standard method for efficiently training machine learning models for data-scarce problems by adding and fine-tuning new classification layers to a model pre-trained on large datasets. Although numerous previous studies proposed to use homomorphic encryption to resolve the data privacy issue in transfer learning in the machine learning as a service setting, most of them only focused on encrypted inference. In this study, we present HETAL, an efficient Homomorphic Encryption based Transfer Learning algorithm, that protects the client's privacy in training tasks by encrypting the client data using the CKKS homomorphic encryption scheme. HETAL is the first practical scheme that strictly provides encrypted training, adopting validation-based early stopping and achieving the accuracy of nonencrypted training. We propose an efficient encrypted matrix multiplication algorithm, which is 1.8 to 323 times faster than prior methods, and a highly precise softmax approximation algorithm with increased coverage. The experimental results for five well-known benchmark datasets show total training times of 567-3442 seconds, which is less than an hour.