Skip to contentScholarGate
LibraryBookshelfDeskReview StudioAssistant
Sign in
On this page
IntuitionHow it worksWhen to use itStrengths & limitationsCommon pitfallsApplicationsFrequently asked🔒 Read the full methodSourcesRelated methods
Cite this pageSpotted an issue on this page? Report or suggest a fix →
Home›Privacy›Homomorphic Encryption
Machine learningPrivacy-preserving computation

Homomorphic Encryption

Fully Homomorphic Encryption · Also known as: FHE, Fully Homomorphic Encryption, Leveled Homomorphic Encryption, Homomorfik Şifreleme

Homomorphic Encryption (HE) is a cryptographic framework that allows arbitrary computations to be performed directly on encrypted data without requiring decryption. First realized as a fully general construction by Craig Gentry in 2009 using ideal lattices, it enables a server to process sensitive data and return an encrypted result that, when decrypted by the data owner, equals the result of performing the same computation on the plaintext. It is foundational to privacy-preserving machine learning, secure cloud computing, and confidential analytics.

ScholarGate
  1. Machine learning
  2. v1
  3. 1 Sources
  4. PUBLISHED
Cite this page →
Tools & resources
Download slides
Learn & explore

Read the full method

Members only

Sign in with a free account to read this section.

Sign in

Method map

The neighbourhood of related methods — select a node to explore.

Homomorphic Encryption
Differential PrivacyFederated LearningSecure Multi-Party Compu…

When to use it

Use homomorphic encryption when data must remain encrypted throughout processing — typically in regulated industries (healthcare, finance, genomics) where a service provider must compute on client data without ever holding plaintext. It is appropriate when the computation can be expressed as an arithmetic or Boolean circuit of bounded or refreshable depth. Avoid it when latency is critical, as even optimized FHE schemes are orders of magnitude slower than plaintext computation. For simpler needs, differential privacy or secure multi-party computation may offer better performance at the cost of different trust models.

Strengths & limitations

Strengths
  • Provides cryptographic guarantees: the server learns nothing about plaintexts, even if fully compromised after computation.
  • Supports arbitrary computable functions over encrypted data via bootstrapping, with no restriction on circuit depth.
  • Compatible with standard public-key infrastructure; the data owner alone holds the secret key.
  • Enables non-interactive privacy-preserving computation, unlike multi-party protocols that require rounds of communication.
Limitations
  • Computational overhead is large: FHE operations can be 10^3 to 10^6 times slower than equivalent plaintext operations depending on scheme and circuit depth.
  • Ciphertext expansion is significant; encrypted data is typically 1000× larger than the corresponding plaintext.
  • Bootstrapping is the dominant cost and remains a practical bottleneck even in hardware-accelerated implementations.
  • Programming FHE circuits requires expert knowledge of noise budgets, parameter selection, and scheme-specific constraints.

Frequently asked

Is fully homomorphic encryption practical today?

For narrow workloads — particularly neural network inference with small models, genomic queries, or simple statistical aggregations — FHE is deployable today using libraries such as SEAL, HElib, Concrete, or OpenFHE. However, latency remains 100× to 10^6× higher than plaintext, so applications must tolerate multi-second or multi-minute runtimes. Hardware acceleration (GPUs, FPGAs, custom ASICs) is an active research area substantially narrowing this gap.

What is the difference between FHE, somewhat homomorphic encryption, and leveled HE?

Somewhat homomorphic encryption (SHE) supports a limited number of multiplications before noise exceeds the decryption threshold. Leveled HE sets parameters large enough to evaluate circuits up to a predetermined depth without bootstrapping. Fully homomorphic encryption adds bootstrapping to refresh noise, supporting circuits of unlimited depth. In practice, many deployments use leveled HE and avoid bootstrapping for performance reasons, accepting a bound on circuit complexity.

Does FHE protect against all privacy attacks?

FHE guarantees that the evaluating party learns nothing from the ciphertexts under standard cryptographic assumptions. However, it does not protect against attacks on the output — a malicious circuit designer can craft computations whose results leak plaintext information. It also does not address membership inference or model inversion at the application layer. FHE is a building block, not a complete privacy solution; it is best combined with output perturbation or differential privacy where output leakage is a concern.

Sources

  1. Gentry, C. (2009). Fully homomorphic encryption using ideal lattices. ACM Symposium on Theory of Computing (STOC), 169–178. DOI: 10.1145/1536414.1536440 ↗

How to cite this page

ScholarGate. (2026, June 2). Fully Homomorphic Encryption. ScholarGate. https://scholargate.app/en/privacy/homomorphic-encryption

Related methods

Differential PrivacyFederated LearningSecure Multi-Party Computation

Which method?

Set this method beside its closest kin and read them side by side — the library lays the books on the table; the choice is yours.

  • Differential PrivacyPrivacy↔ compare
  • Federated LearningPrivacy↔ compare
  • Secure Multi-Party ComputationPrivacy↔ compare
Compare side by side →

Similar methods

Secure Multi-Party ComputationDifferential PrivacyLattice-Based Cryptographyzk-SNARKFederated LearningPost-Quantum Cryptography (Kyber)Zero-Knowledge Proofzk-STARK

Related reference concepts

Secure Multiparty ComputationDe-identification and Privacy-Preserving Data AnalysisComputational Hardness AssumptionsPublic-Key CryptographyPost-Quantum CryptographyZero-Knowledge Proofs

Spotted an issue on this page? Report or suggest a fix →

ScholarGate — Homomorphic Encryption (Fully Homomorphic Encryption). Retrieved 2026-07-21 from https://scholargate.app/en/privacy/homomorphic-encryption · Dataset: https://doi.org/10.5281/zenodo.20539026
Quick facts
Originator
Craig Gentry
Year
2009
Type
Lattice-based cryptographic scheme
Subfamily
Privacy-preserving computation
Hardness Assumption
Learning With Errors (LWE) / ideal lattices
Operation Classes
Addition and multiplication over ciphertexts
Related methods
Differential PrivacyFederated LearningSecure Multi-Party Computation
ScholarGate

A content-first reference library for research methods — what each one is, how it works, and where it comes from.

Open data (CC-BY)

Explore

  • Library
  • Search the library…
  • Browse by field
  • Fields
  • Journey
  • Compare
  • Which method?

Reference

  • Subjects
  • Atlas
  • Glossary
  • Methodology
  • Philosophy

Your tools

  • Bookshelf
  • Desk
  • Chat

Company

  • About
  • Pricing
  • Contact
  • Suggest a method

Entries are compiled from published sources for reference. Verifying the accuracy and suitability of any information for your own use remains your responsibility.

© 2026 ScholarGate · A research-method reference library
  • Privacy
  • Cookies
  • Terms
  • Delete account