Skip to content

Private svm inference on encrypted data

Authors: A. Al Badawi
Year: 2024

type: textbook collection: sok-references status: auto

Links: link

Summary

Auto-extracted summary (unreviewed)

This tutorial book chapter walks through building privacy-preserving SVM inference with the CKKS FHE scheme in a machine-learning-as-a-service setting, covering data preparation, scikit-learn training and homomorphic inference via the FHSVM engine on OpenFHE. On a credit card fraud detection dataset it reports high predictive accuracy with an encrypted inference latency of about 6 seconds on a commodity laptop.

Abstract

This tutorial chapter provides a comprehensive guide to implementing privacypreserving Support Vector Machine (SVM) inference using Fully Homomorphic Encryption (FHE). We demonstrate a practical solution for secure and private SVM inference on encrypted data, enabling sensitive data analysis while maintaining confidentiality. Through a step-by-step implementation on a real-world dataset, we cover data preparation, SVM model training, and homomorphic inference. Our experimental results on a commodity laptop show that our approach achieves high accuracy with a reasonable latency of nearly 6 seconds for SVM inference. This chapter serves as a valuable resource for practitioners and researchers seeking to apply privacypreserving techniques to SVM solutions, with significant implications for applications like medical diagnosis, financial prediction, and recommender systems, where data privacy is crucial. By following this tutorial, readers can gain hands-on experience with privacy-preserving SVM inference using FHE.
Source: pdf

Extracted details

Field Value Source Confidence
contribution step-by-step tutorial for CKKS-based SVM inference on encrypted data llm 0.80
contribution end-to-end pipeline: scikit-learn training, model extraction, FHSVM/OpenFHE homomorphic inference llm 0.80
contribution case study on credit card fraud detection with ~6 s encrypted inference llm 0.80
schemes CKKS llm 0.80
libraries OpenFHE llm 0.80
libraries FHSVM llm 0.80
libraries scikit-learn llm 0.80
models SVM llm 0.80
datasets credit card fraud (Dal's dataset) llm 0.80
tasks fraud detection llm 0.80
tasks binary classification llm 0.80
tasks machine learning as a service llm 0.80
hardware commodity laptop llm 0.80
threat model untrusted server, client-server (MLaaS) llm 0.80
interaction non-interactive FHE llm 0.80
limitations future work: optimising homomorphic SVM inference performance and exploring alternative FHE schemes llm 0.80
llm relevance low llm 0.80
key result inference latency: ~6 s (CKKS SVM inference on a commodity laptop, credit card fraud dataset) llm 0.80

SoK reference [9]

A. Al Badawi, “Private svm inference on encrypted data,” in Federated Learning-A Systematic Review. IntechOpen, 2024.

Listed on the reference page of the site.

Keyword mentions

Counts of vocabulary terms in the full text (first page of occurrence in brackets).

  • scheme: CKKS 49 [p.3], MPC 2 [p.6]
  • library: OpenFHE 8 [p.9]
  • model: SVM 114 [p.1], CNN 4 [p.3], Transformer 3 [p.3], Logistic regression 3 [p.3]
  • dataset: Credit card / German credit 8 [p.8]
  • technique: Federated learning 10 [p.2], Semi-honest model 8 [p.3], Rotations 7 [p.4], Polynomial approximation 5 [p.7], Bootstrapping 4 [p.4], Sigmoid 3 [p.6], Encrypted training 2 [p.19]
  • assumption: Lattice-based 2 [p.19]
  • task: Medical / healthcare 6 [p.1], Genomics 6 [p.6], Finance 3 [p.1]
  • topic: Homomorphic encryption 60 [p.1], Open source 4 [p.6], Privacy-preserving machine learning 3 [p.8]

Files

Path (under papers/) Collection Pages Size SHA-256 Note
references/sok-references/009_private-svm-inference-on-encrypted-data.pdf sok-references 20 2.0 MB 5943bdfb78c8

Database key: private-svm-inference-on-encrypted-data · created 2026-11-05 · updated 2026-11-05 · edit with python3 tools/paperdb.py set private-svm-inference-on-encrypted-data COLUMN VALUE