mlsanityData to 14 Oct 2026

Gecko: Fast Private Inference via Secure Public Encoder Offloading

arXiv 2608.02378 ↗First posted 3 Aug 2026cs.LGcited by 0

Not on the map. Why?

Abstract (arXiv)

Private inference protects both user inputs and server models during neural network inference, but existing solutions remain too slow for practical deployment. This motivates recent efforts to run a public encoder, such as a pretrained backbone, outside the protection boundary and evaluate only a small private predictor cryptographically. While appealing for efficiency, this design is not inherently secure: naively offloading a public encoder may create a feature-space shortcut: an extraction adversary may learn the remaining private predictor's feature-to-output mapping more easily than the original model's input-to-output behavior. We present Gecko, designed to limit this additional risk while retaining a compact encrypted predictor. We leverage a frozen backbone that contributes hierarchical features, fixed Fastfood projections that compress them, and private feature gating that prepares them for prediction. We formalize ideal independence and information-preservation conditions as design guidance, then separately evaluate component-reuse extraction attacks. Across image and audio tasks, Gecko achieves 0.4-2.2 second inference with at most 10.8 MB communication and accuracy comparable to transfer-learning baselines. Under the evaluated attacks, reusing the offloaded public encoder provides no significant advantage to model-extraction adversaries. Source code and a demo are available at https://github.com/CassiniHuy/gecko-infer.

Key facts. No citing papers in the map yet (data through 14 Oct 2026). Builds on 32 of its 63 references that are in the map. First posted on arXiv 3 Aug 2026.

Builds on

32 of 63 references.

References · 32

  1. 2015Deep Residual Learning for Image Recognitioncited by 52,096
  2. 2014ImageNet Large Scale Visual Recognition Challengecited by 8,498
  3. 2019Searching for MobileNetV3cited by 1,994
  4. 2016Identity Mappings in Deep Residual Networkscited by 1,965
  5. 2017EuroSAT: A Novel Dataset and Deep Learning Benchmark for Land Use and Land Cover Classificationcited by 1,496
  6. 2019A Comprehensive Survey on Transfer Learningcited by 1,160
  7. 2021MedMNIST v2 -- A large-scale lightweight benchmark for 2D and 3D biomedical image classificationcited by 690
  8. 2017Opening the Black Box of Deep Neural Networks via Informationcited by 503
  9. 2018Knockoff Nets: Stealing Functionality of Black-Box Modelscited by 292
  10. 2015Quantized Convolutional Neural Networks for Mobile Devicescited by 215
  11. 2020MedMNIST Classification Decathlon: A Lightweight AutoML Benchmark for Medical Image Analysiscited by 206
  12. 2015Structured Pruning of Deep Convolutional Neural Networkscited by 149
20 more
  1. 2019Understanding the Limitations of Variational Mutual Information Estimatorscited by 140
  2. 2022Head2Toe: Utilizing Intermediate Representations for Better Transfer Learningcited by 83
  3. 2014Fastfood: Approximate Kernel Expansions in Loglinear Timecited by 74
  4. 2020DarkneTZ: Towards Model Privacy at the Edge using Trusted Execution Environmentscited by 58
  5. 2018Low Latency Privacy Preserving Inferencecited by 57
  6. 2022THE-X: Privacy-Preserving Transformer Inference with Homomorphic Encryptioncited by 57
  7. 2022Visual Query Tuning: Towards Effective Usage of Intermediate Representations for Parameter and Memory Efficient Transfer Learningcited by 57
  8. 2017CryptoDL: Deep Neural Networks over Encrypted Datacited by 56
  9. 2023Understanding Deep Representation Learning via Layerwise Feature Compression and Discriminationcited by 32
  10. 2023No Privacy Left Outside: On the (In-)Security of TEE-Shielded DNN Partition for On-Device MLcited by 21
  11. 2023Converting Transformers to Polynomial Form for Secure Inference Over Homomorphic Encryptioncited by 19
  12. 2023AutoFHE: Automated Adaption of CNNs for Efficient Evaluation over FHEcited by 15
  13. 2022Private, Efficient, and Accurate: Protecting Models Trained by Multi-party Learning with Differential Privacycited by 7
  14. 2024PrivCirNet: Efficient Private Inference via Block Circulant Transformationcited by 7
  15. 2025Breaking the Layer Barrier: Remodeling Private Transformer Inference with Hybrid CKKS and MPCcited by 7
  16. 2023NeuJeans: Private Neural Network Inference with Joint Optimization of Convolution and FHE Bootstrappingcited by 6
  17. 2023Fast and Private Inference of Deep Neural Networks by Co-designing Activation Functionscited by 4
  18. 2025Comet: Accelerating Private Inference for Large Language Model by Predicting Activation Sparsitycited by 4
  19. 2025CryptPEFT: Efficient and Private Neural Network Inference via Parameter-Efficient Fine-Tuningcited by 2
  20. 2026PrivDNN: A Secure Multi-Party Computation Framework for Deep Learning using Partial DNN Encryptioncited by 1

Built on by

Not cited yet.