Karim Eldefrawy

Cryptography, Cybersecurity, Privacy

Co-founder and CTO at Confidencial.io
2017-2021: SRI
2011-2016: HRL Laboratories
2006-2010: PhD@UC Irvine

Scientific curiosity

Scientific knowledge map · Paper #65

On the Hardness of Scheme-Switching Between SIMD FHE Schemes

Karim Eldefrawy, Nicholas Genise, and Nathan Manohar

2023 · 14th International Conference on Post-Quantum Cryptography (PQCrypto)

  • Theory

What does the paper try to establish?

Can exact SIMD ciphertexts from BGV or BFV be efficiently converted to approximate CKKS ciphertexts, and back, without making expensive FHE capabilities such as bootstrapping unexpectedly cheap?

What is the proposed answer?

The paper gives conditional reductions rather than an unconditional lower bound: sufficiently capable scheme-switching oracles imply bootstrapping for the target schemes with little additional homomorphic work. It also relates comparison access to bootstrapping and nonlinear functions, supplying barrier evidence for general-purpose switching across exact and approximate encodings.

Six dimensions, kept separate

The chart summarizes documented evidence and process. It is not a correctness probability, confidence score, or ranking, and no composite score is calculated.

The visual spider chart requires JavaScript. The complete values and rationales follow in text.

LowMediumHighN/A = not assessed

A smaller value means less documented support for that dimension, not that the paper is false or unimportant.

Epistemic evidence High

The full source supplies formal oracle interfaces, lemmas, reductions, and parameter conditions. High describes the documented support for conditional theorems, not an unconditional hardness proof.

Weak and strong scheme-switch definitions BGV-to-CKKS switching implies CKKS bootstrapping CKKS-to-BGV switching implies BGV bootstrapping
Auditability High

The full IACR manuscript and author record make assumptions, theorem statements, and proofs publicly inspectable, while the official DOI fixes the publication identity.

Problem statement, contributions, and interpretation Official publication identity
Production provenance Medium

Named authors, public manuscript, venue, year, and DOI are documented; author roles, revision history, tools, and proof-development process are not.

Problem statement, contributions, and interpretation Official publication identity
External scrutiny Medium

PQCrypto publication establishes venue scrutiny, but no public reports, independent proof audit, formal verification, or correction history was located.

Official publication identity
Reception Low

The dated OpenAlex snapshot located 1 citation. Under the author-defined rule, 0 through 8 located citations is Low; counts do not measure validity.

Dated citation-count snapshot
Contribution significance High

The reductions expose a general barrier between scheme switching and bootstrapping across major SIMD FHE families. This is a significance assessment of the theorem's scope, not a claim of unconditional impossibility.

Problem statement, contributions, and interpretation BGV-to-CKKS switching implies CKKS bootstrapping CKKS-to-BGV switching implies BGV bootstrapping

Assessment: Ai draft author review pending · 2026-07-11 · rubric 0.2. These dimensions describe documented support and process, not truth, correctness, or a universal ranking. No composite score is calculated.

Hierarchical knowledge map

Collapse a branch for a top-level reading, or follow its source links and child nodes to audit the evidence and boundaries underneath it.

paper

Hardness of SIMD FHE scheme switching

A theory paper that studies exact-to-approximate and approximate-to-exact SIMD FHE conversion through oracle reductions, showing that broadly useful switching would subsume costly refresh capabilities.

Problem statement, contributions, and interpretation
  1. model Exact and approximate SIMD model defined

    BGV and BFV represent exact modular plaintext slots, whereas CKKS represents approximate complex or real slots with scale and error; switching must address both ciphertext semantics and plaintext encoding.

    BGV, BFV, CKKS, SIMD encodings, and oracle notation
    1. definition

      Weak versus strong switching

      formalized

      A weak oracle preserves a compatible underlying ring-polynomial representation without performing the full slot-encoding conversion. A strong oracle includes the encoding transformation expected by an application-facing scheme switch.

      Weak and strong scheme-switch definitions
    2. theorem

      Weak-to-strong lifting

      proved

      The Section 3 lemmas and theorems connect the weak oracle formulations to strong switching by accounting for the relevant BGV and CKKS encoding transformations.

      Weak and strong scheme-switch definitions
  2. claim group Reduction results proved

    The central results are black-box consequences of possessing the specified switching oracle; they are not constructions of a cheap switching implementation.

    BGV-to-CKKS switching implies CKKS bootstrapping CKKS-to-BGV switching implies BGV bootstrapping
    1. theorem

      Exact-to-approximate switch bootstraps CKKS

      proved

      Theorem 5 shows that a weak BGV-to-CKKS switching oracle can be used to bootstrap CKKS with one oracle call and one rescaling step, subject to the theorem's parameter and noise conditions.

      BGV-to-CKKS switching implies CKKS bootstrapping
    2. theorem

      Approximate-to-exact switch bootstraps BGV

      proved

      The reverse reduction uses one CKKS-to-BGV oracle call followed by a homomorphic inner-product style computation, plaintext multiplication and addition, and modulus switching to obtain a refreshed BGV ciphertext; the treatment extends to BFV through the stated conversion.

      CKKS-to-BGV switching implies BGV bootstrapping
    3. theorem

      Comparison access also carries refresh power

      proved conditionally

      The comparison analysis shows how a suitable comparison oracle can support bootstrapping with few calls for typical parameter regimes and connects the mechanism to nonlinear operations such as minimum, maximum, and ReLU.

      Comparison-oracle consequences
  3. limitation group What the reductions do not establish explicitly bounded

    The reductions delimit the consequences of modeled switching capabilities but do not prove an unconditional complexity lower bound or impossibility theorem.

    Problem statement, contributions, and interpretation Interpretation and open scope
  4. scrutiny

    External scrutiny

    venue reviewed

    PQCrypto publication establishes venue review. No public proof review, correction, independent formalization, or subsequent technical critique was located in this audit.

    Official publication identity

Source index

Locators state the depth of the current audit. PDF page numbers, where present, are one-based file pages; metadata-, summary-, and abstract-bounded records explicitly identify their limitations.

  1. Problem statement, contributions, and interpretation Abstract and Section 1
  2. BGV, BFV, CKKS, SIMD encodings, and oracle notation Section 2
  3. Weak and strong scheme-switch definitions Section 3, including Lemmas 3.1-3.2 and Theorems 3-4
  4. BGV-to-CKKS switching implies CKKS bootstrapping Section 4, Theorem 5
  5. CKKS-to-BGV switching implies BGV bootstrapping Section 4, Theorem 6 and BFV extension
  6. Comparison-oracle consequences Section 5
  7. Interpretation and open scope Conclusion
  8. Official publication identity PQCrypto 2023, DOI 10.1007/978-3-031-40003-2_8
  9. Dated citation-count snapshot OpenAlex cited_by_count was 1 when accessed 2026-07-11